Review Article Beyond DNA: An Integrated and Functional …downloads.hindawi.com/journals/ijbc/2016/2469523.pdf · 2019-07-30 · Review Article Beyond DNA: An Integrated and Functional
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Review ArticleBeyond DNA An Integrated and Functional Approach forClassifying Germline Variants in Breast Cancer Genes
T Pesaran R Karam R Huether S Li S Farber-Katz A ChamberlinH Chong H LaDuca and A Elliott
Ambry Genetics Corp 15 Argonaut Aliso Viejo CA 92656 USA
Correspondence should be addressed to T Pesaran tpesaranambrygencom
Received 4 June 2016 Revised 4 September 2016 Accepted 19 September 2016
Academic Editor Zsuzsanna Kahan
Copyright copy 2016 T Pesaran et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
Genetic testing for hereditary breast cancer is an integral part of individualized care in the new era of precision medicineThe accuracy of an assay is reliant on not only the technology and bioinformatics analysis utilized but also the experience andinfrastructure required to correctly classify genetic variants as disease-causing Interpreting the clinical significance of germlinevariants identified by hereditary cancer testing is complex and has a significant impact on the management of patients who are atincreased cancer risk In this review we give an overview of our clinical laboratoryrsquos integrated approach to variant assessment Wediscuss some of the nuances that should be considered in the assessment of genomic variants In addition we highlight lines ofevidence such as functional assays and structural analysis that can be useful in the assessment of rare and complex variants
1 Introduction
The landscape of genetic testing for hereditary breast cancersusceptibility has changed drastically with the application ofmassively parallel sequencing based tests in clinical diag-nostics Clinical genomic laboratories are performing anincreasing number of massively parallel sequencing assaysfor cancer predisposition genes [1] which has led to anintensified application of these assays in clinical and researchsettings [2] Breast cancer gene panels and exome sequencinggenerate vast amounts of genetic alteration data therebypresenting a significant challenge to determine which vari-ants are responsible for the disease or phenotype Multigenebreast cancer panels in particular have gained in popularityover the past few years and are now routinely ordered bygenetics oncology and breast surgical clinics These testsallow for simultaneous analysis of numerous cancer genesthat when mutated can have a significant impact on cancerrisk stratification and management [3] A major componentof clinical molecular diagnostic testing is accurate assessmentand interpretation of genetic variants
MRE11A MUTYH NBN NF1 PALB2 PTEN RAD50RAD51C RAD51D and TP53) by massively parallel sequenc-ing of all coding exons and aminimumof 5 base pairs into theflanking 51015840 and 31015840 ends of all introns and untranslated regionsIn addition clinically significant intronic mutations beyond5 base pairs and the promoter region of PTEN (c-1300 to c-745) are always sequenced and reported Sequencing is con-ducted on the Illumina HiSeq2500 or NextSeq using 150 bppaired-end conditions as described in the manufacturerrsquosstandard workflow (Illumina) After initial data processingall clinical samples had to pass minimum thresholds tobe included in the analysis The three parameters were asfollows mean base calling quality score is greater than 30the percentage of passes that reached over 30 had to be75 overall and the percentage of perfectly matched indexesneeded to be greater than 85 For each gene a minimumcoverage of 20x is required for candidate variants to be called
In an effort to help standardize the interpretation andreporting of genetic testing results organizations such asthe American College of Medical Genetics and Genomics(ACMG) Association for Molecular Pathology (AMP) andthe International Agency for Research and Cancer (IARC)have proposed criteria for the interpretation and reporting of
Hindawi Publishing CorporationInternational Journal of Breast CancerVolume 2016 Article ID 2469523 10 pageshttpdxdoiorg10115520162469523
2 International Journal of Breast Cancer
sequence variants [7ndash9]These criteria weighmultiple lines ofevidence to categorize variants under a five-tier classificationalgorithm using terms such as pathogenic (P) variant likelypathogenic (VLP) variant of unknown significance (VUS)variant likely benign (VLB) and benign (B) to indicate thelikelihood of association with disease Per ACMG guidelinesthe term ldquolikelyrdquo refers to a classification tier that equates to agt90 likelihood of a variant being disease-causing or benign[7 8] Recently the clinical utility of the ACMG guidelineswas demonstrated in a cohort of individuals undergoingsequencing for inherited cancer risk [10]
While the ACMG guidelines provide a basic frameworkfor variant assessment gene and syndrome-specific factorssuch as penetrance prevalence inheritance pattern diseasemechanism and protein structure and function need to beconsidered In addition when considering the phenotype ofthe patients inwhich a variant is identified onemust take intoaccount the prevalence of the disease and how the patientsare ascertained to account for potential phenocopies Forexample many genes on hereditary breast cancer panels areconsidered to be moderate penetrance and are associatedwith a 2- to 5-fold increased breast cancer risk Given therelatively high prevalence of breast cancer (18 women inthe US) traditional segregation methods are confounded byphenocopies and are evenmore difficult to employwith genesthat have reduced penetrance These confounders indicatethat these genes require large numbers of segregation eventsto provide meaningful results Consideration should also begiven to gene-specific factors such as frequency of germlineand somatic de novo alterations additional tests in tumorssuch as loss of heterozygosity studies variation in nonsense-mediated decay and alternate splicing For example in genessuch asTP53 and PTEN germline de novo variants are knownto be a relatively common cause of disease [11 12] Howeverwith breast cancer genes such as ATM CHEK2 and PALB2the de novo rate is unknown This is confounded by the factthat breast cancer is a common disease and one cannot inferif the de novo event in these genes directly correlates withdisease or occurred by chance In addition although somaticde novo data is available for some genes [13] its incorporationinto germline variant analysis has yet to be standardized andwill need to be performed on a gene-by-gene basis
Consortia such as the Evidence Based Network forthe Interpretation of Germline Mutant Alleles (ENIGMA)have demonstrated the power of a collaborative approachto variant assessment and have made great strides in thereclassification of VUS in breast cancer genes as pathogenicor benign However even these groups are limited by the rateat which data is accumulated Open-access databases suchas ClinVar and the Leiden Open (source) Variant Database(LOVD) can be useful in identifying additional cases orpublications related to a variant These databases have alsohelped standardize the interpretation of variants betweenlaboratories by identifying discrepancies in classificationsCollaborative efforts by clinical laboratories includingAmbryGenetics GeneDx University of Chicago and Laboratoryfor Molecular Medicine have resulted in the sharing ofinternal data consisting of segregation and cooccurrenceswith mutations in the same gene or other genes and de novo
observations have led to the resolution of 78 of clinicallyactionable differences (VUS versus VLPmutation) and 92of VUS versus likely benignbenign differences (internaldata) Despite these efforts one of the challenges faced bymolecular laboratories and clinicians is that many geneticvariants are very rare and do not have enough published datato be classified beyondVUSWe present here our laboratoryrsquosintegrated approach to variant assessment and review toolsused to assess the impact of variants on protein function
2 Integrated Approach to Variant Assessment
Ambry Genetics has developed and implemented an inte-grated approach to variant assessment (Table 1) that encom-passes a five-tier variant classification algorithm similar tothose presented by ACMG and IARC Although the founda-tion of Ambry Geneticsrsquo classification algorithm is based onthe ACMG guidelines we have adopted stringent thresholdssimilar to those proposed by the IARC where ldquolikelyrdquo refersto a gt95 confidence of a variant being disease-causing orbenign [9] In this algorithm both pathogenic and likelypathogenic variants are interpreted as clinically actionablewith recommendations for medical management and familymember testing
Ambry Geneticsrsquo algorithm incorporates multiple lines ofevidence aimed at assessing both the impact of the variant onthe protein and the pathogenicity of the variant in relationto a disease phenotype (Figure 1 and Table 2) These lines ofevidence are weighted as stand-alone (categories A and F)strong (categories B and D) or supportive (categories C andE) and when combined as described in Table 1 they can leadto a classification of likely benign benign likely pathogenicor pathogenic When the evidence is limited or conflictingthe variants remain classified as VUS Lines of evidence suchas its location structure-function and functional and RNAstudies reflect the functional impact on themRNAor proteinEvolutionary conservation in silicomodels such as Polyphenand SIFT and general population frequency reflect fitnessthat is reproductive success and survival as measured by alack of allelic diversity The observed phenotype in variantcarriers and the cosegregation of the variant with disease andthe cooccurrence with other pathogenic variants reflect thepathogenicity of the variant (Figure 1) Some of this evidenceis readily available via databases such as allele frequencydata in the Exome Aggregation Consortium (ExAC) or thedata in published literature [14] However published literaturegenerally contains data for common variants and the datasupporting pathogenicity for rare variants is scarce andfrequently only available internally
For most genes on breast cancer panels computationaldata from in silico models evolutionary conservation andprotein structural analysis are readily available Populationfrequency data has been accumulating at a fast pace dueto major contributions from 1000 Genomes NHLBI ExomeSequencing Project (ESP) and ExAC These data have hada significant impact on the identification of benign variantsat high frequencies that are too frequent to be pathogenicbased on disease incidence alone particularly for historicallyunderstudied ethnic groups For breast cancer genes this
International Journal of Breast Cancer 3
Table1Classifi
catio
nschemefor
high
penetrance
autosomaldo
minantb
reastcancerg
enes
Class
Classifi
catio
nCa
tegory
Criteria
5Pathogenic
A1n
eeded
(stand
-alone)
(i)Alteratio
nsresulting
inprem
aturetruncation(egreadingfram
eshift
nonsense)
(ii)O
ther
ACMG-definedmutations
(ie
initiationcodo
nor
grossd
eletion)
(iii)Strong
segregationwith
disease(LO
Dgt3=gt10
meioses)
5Pathogenic
B4needed
(stro
ng)
(i)Con
firmed
denovo
alteratio
nin
thes
ettin
gof
anew
disease(approp
riateph
enotype)in
thefam
ily(ii)S
ignificantd
iseasea
ssociatio
nin
approp
riatelysiz
edcase-con
trolstudy(ie
s)(iii)Be
ingdetected
inindividu
alssatisfying
establish
eddiagno
sticc
riteriaforc
lassicdiseasew
ithou
tacle
armutation
(iv)L
astn
ucleotideo
fexon
(v)G
oodsegregationwith
disease(LO
D15
ndash3=5ndash9meioses)
(vi)Deficientp
rotein
functio
nin
approp
riatefunctio
nalassay(s)
(vii)
Functio
nally
valid
ated
splicingmutation
(viii)W
ell-characterized
mutationatthes
amep
osition
(ix)O
ther
stron
gdatasupp
ortin
gpathogeniccla
ssificatio
n(egstr
uctural)
4Likelypathogenic
1needed
(i)Alteratio
nsatthec
anon
icaldo
noracceptor
sites
(plusmn12)
with
outano
ther
stron
g(B-le
vel)evidence
supp
ortin
gpathogenicity
4Likelypathogenic
C4needed
(sup
portive)
(i)Ra
rityin
generalp
opulationdatabases(db
SNPES
P1000
Genom
esE
xAC)
(ii)Insilico
mod
elsinagreem
ent(deleterio
us)a
ndorc
ompletely
conservedpo
sitionin
approp
riatespecies
(iii)Mod
erates
egregatio
nwith
disease(atleast3
inform
ativem
eioses)for
rare
diseases
(iv)O
ther
datasupp
ortin
gpathogeniccla
ssificatio
n(egstr
uctural)
3of
B2of
Bandatleast1
ofC
1ofB
andatleast3
ofC
3VUS
Insufficiento
rcon
flictingevidence
Grossdu
plications
with
outstro
ngevidence
forp
atho
genico
rbenign
3Likelybenign
D1n
eeded
(stro
ng)
(i)Intro
nica
lteratio
nwith
nosplicingim
pactby
RT-PCR
analysisor
anothersplicingassay
(ii)O
ther
stron
gdatasupp
ortin
gbenign
classificatio
n
3Likelybenign
E2needed
(sup
portive)
(i)Coo
ccurrences
with
mutations
inthes
ameg
ene(ph
aseu
nkno
wn)
(ii)C
ooccurrences
with
mutations
inotherh
ighpenetra
ntgenesthatclearlyexplainap
roband
rsquosph
enotype
(iii)Subp
opulationfre
quency
insupp
orto
fbenigncla
ssificatio
n(iv
)Intactp
rotein
functio
nob
served
inapprop
riatefunctio
nalassay(s)
(v)Insilico
mod
elsinagreem
ent(benign
)(vi)Not
segregatingwith
diseaseinfamily
study
(genes
with
incompletep
enetrance)
(vii)
Nodiseasea
ssociatio
nin
smallcase-controlstudy
(viii)O
ther
datasupp
ortin
gbenign
classificatio
n
1Be
nign
F1n
eeded
(stand
-alone)
(i)Generalpo
pulationor
subp
opulationfre
quency
istoohigh
tobe
apatho
genicm
utationbasedon
diseasesynd
romep
revalencea
ndpenetrance
(ii)N
otsegregatingwith
diseaseinfamily
study
(genes
with
completep
enetrance)
(iii)Internalfre
quency
istoohigh
tobe
apatho
genicm
utationbasedon
diseasesynd
romep
revalencea
ndpenetra
nce
(iv)B
eing
seen
intra
nswith
amutationor
inho
mozygou
sstatein
individu
alsw
ithou
tsevered
iseasefor
thatgene
(v)N
odiseasea
ssociatio
nin
approp
riatelysiz
edcase-con
trolstudy(ie
s)1o
fDandatleast2
ofE
2or
moreo
fDgt3of
Ewo
confl
ictin
gdata
gt4of
Ewcon
flictingdata
Thev
ariant
classificatio
nschemeisn
otintend
edforthe
interpretatio
nof
alteratio
nscomplicated
byepigeneticfactorsincluding
genetic
mod
ifiersmultifactoria
ldise
aseor
low-risk
diseasea
ssociatio
nallelesa
ndmay
belim
itedin
theinterpretationof
alteratio
nsconfou
nded
byincompletep
enetrancevaria
blee
xpressivitypheno
copiesand
triallelic
oroligogenicinheritance
4 International Journal of Breast Cancer
Variantclassification
Fitness
Patho
genicity
Function
In silicoassessment
Polyphen SIFTmutationAssessor
Provean ESE fruitfly
Evolutionaryconservation
UCSC genome browserNCBI
Populationfrequency data
1000 Genomes ESPExAC
Case-controlstudies
Published literatureinternal data
Phenotype
Published literatureinternal data
Cooccurrenceamp segregation
Published literatureinternal data
Structural function
Uniprot wwPDBinternal dataRNA studies
Published literatureinternal data
Functional assays
Published literatureinternal data
Figure 1 An integrated approach for variant classification Lines of evidence such as structural function RNA studies and functional studiesassess the functional impact on the mRNA and protein Cooccurrence segregation case-control studies and the observed phenotype invariant carriers reflect the pathogenicity of a variant Population frequency in silico models and evolutionary conservation assess fitness ofthe amino acid or nucleotide position
threshold has been conservatively set at an allele frequencyof 1 in large population cohorts if used as a stand-alone lineof evidence supporting benign classification (Table 1 categoryF) Careful consideration of population cohort size is neededto attain a high confidence (lower 95 CI is above 1 with119901 value lt005) that the frequency is above 1 For examplewith a cohort of 60000 alleles an allele frequency of 108is sufficient (lower 95 CI = 101 119901 = 00244) whereas fora cohort of 1000 alleles an allele frequency of 170 (lower95 CI = 115 119901 = 0013) is needed to be 95 confidentthe allele frequency is above 1
Although patient phenotype cooccurrence and cosegre-gation data can be found in the published literature manylaboratories also curate internal data for use in variant classi-fication A patientrsquos clinical and family history can be difficultto use as a line of evidence in a clinical laboratory settingdue to ascertainment bias However when a variant in a geneassociated with a rare disorder (less than 12000) is identifiedin multiple individuals meeting classic clinical criteria andnever in large control populations or population cohortsthese data can be used as evidence towards pathogenicityThis is most informative in patients who have undergonegenetic testing on large multigene panel tests in which allthe known genes associated with a disorder have been ruledout However when defining classic clinical criteria we usevery strict guidelines and exclude common diseases such as
breast cancer For example when assessing a TP53 variantthe phenotype is considered strong if the patientmeets classicLi-Fraumeni syndrome criteria a proband with sarcomadiagnosed before 45 years a first-degree relative with anycancer before 45 years and a second-degree relative with anycancer before age 45 years or a sarcoma at any age [15] Forcommon diseases and moderate penetrance genes Bayesiananalyses that require larger phenotype data sets are used [16]Historically in vitro studies were predominantly found in thepublished literature However due to the rapid accumulationof rare variants clinical laboratories such as Ambry Geneticsare implementing validated internal functional studies suchas splicing and homology-directed DNA break repair (HDR)assays that can be incorporated into variant classificationalgorithms
3 Functional Lab
Many variants are classified as VUS because their functionalimpact either is poorly understood or has not yet beeninvestigated These variants include missense and splicingalterations in tumor suppressor genes that require loss offunction to manifest a disease [7] Clinical genomic labora-tories have traditionally relied on evidence from publishedliterature to establish the impact of a variant on gene expres-sion or protein function [7] There are several limitations
International Journal of Breast Cancer 5
DNA-NGS test
Detection of a VUS
Functional lab
Missense VUS Splicing VUS Gross duplication
Functional assays
For example homologousrecombination
RNA studies
For example CE cloning andsequence analysis
Breakpoint studies
Analysis of insertionbreakpoints
Figure 2 Workflow of a functional lab for the evaluation of VUS
Table 2 Experimental structures of genes linked to breast cancerlowast
Gene Length PDBs Coverage ()ATM 3056 0 00BARD1 777 5 421BRCA1 1863 27 176BRCA2 3418 2 16BRIP1 1249 3 19CDH1 882 12 262CHEK2 543 38 864MRE11A 708 1 581MUTYH 549 2 773NBN 754 0 00NF1 2839 6 221PALB2 1186 2 297PTEN 403 6 928RAD50 1312 0 00RAD51C 376 0 00RAD51D 328 1 253TP53 393 142 1000lowastGene lengths and coverage are tabulated from the Universal ProteinResource (Uniprot) [4] and the Research Collaboratory for StructuralBioinformatics (RCSB) [5] databases The list of genes is taken from theBreastNext panel
to this approach including publication bias difficulties withpromptly obtaining additional information about results andprotocols and lack of published evidence for a specificalteration One potential solution is for clinical genomiclaboratories to implement a ldquofunctional labrdquo that can generateassays with high sensitivity and specificity (gt99) and pro-vide unbiased molecular evidence to elucidate the functionalimpact of a VUS (Figure 2) As an example of a convincinglyvalidated assay Guidugli and colleagues determined the
sensitivity of their homology-directed DNA break repair(HDR) functional assay to be 100 (95 confidence interval(CI) 753ndash100) and the specificity to be 100 (95 CI815ndash100) [17]
31 RNA Studies for Splicing VUS While some splicing vari-ants such as canonical plusmn1 or 2 splice sites are often assumedto disrupt gene function by leading to the reduced expressionof the abnormal allele due to nonsense-mediated decay(NMD) [18] or abnormal protein truncations [19] compre-hensive evaluation of splicing alterations is essential for accu-rate clinical interpretation For canonical splice site plusmn1 and 2variants onemust also consider the possibility of an in-framedeletioninsertion which could retain the critical regions ofthe protein and hence lead to a mild neutral or gain-of-function effect In addition variants that are predicted toimpact splicing but that are not located at the canonical sites(plusmn1 and 2) require additional strong evidence (see details inSection 2) to be classified as pathogenic or benign [7] Bioin-formatics software has been developed to predict putativesplice sites [20] In general these in silico tools are more sen-sitive (sim90ndash100) than being specific (sim60ndash80) when pre-dicting the impact of a variant on splicing [21 22] Howeverby nature in silico tools can only provide supporting evidencewhich restricts their use [7] Consequently data from RNAsplicing assays designed to provide quantitative and qualita-tive characterization of transcripts are usually necessary toevaluate the pathogenicity of these variants Since publishedRNA data is not available for every variant clinical genomiclaboratories can more accurately classify splicing alterationsby implementing their own RNA protocols and assays toprovide accurate classification of splicing alterations
Reliability in which an assay yields the same results inrepeated trials is a key issue when implementing mRNAassays in a clinical functional lab for evaluation of VUS
6 International Journal of Breast Cancer
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
sequence variants [7ndash9]These criteria weighmultiple lines ofevidence to categorize variants under a five-tier classificationalgorithm using terms such as pathogenic (P) variant likelypathogenic (VLP) variant of unknown significance (VUS)variant likely benign (VLB) and benign (B) to indicate thelikelihood of association with disease Per ACMG guidelinesthe term ldquolikelyrdquo refers to a classification tier that equates to agt90 likelihood of a variant being disease-causing or benign[7 8] Recently the clinical utility of the ACMG guidelineswas demonstrated in a cohort of individuals undergoingsequencing for inherited cancer risk [10]
While the ACMG guidelines provide a basic frameworkfor variant assessment gene and syndrome-specific factorssuch as penetrance prevalence inheritance pattern diseasemechanism and protein structure and function need to beconsidered In addition when considering the phenotype ofthe patients inwhich a variant is identified onemust take intoaccount the prevalence of the disease and how the patientsare ascertained to account for potential phenocopies Forexample many genes on hereditary breast cancer panels areconsidered to be moderate penetrance and are associatedwith a 2- to 5-fold increased breast cancer risk Given therelatively high prevalence of breast cancer (18 women inthe US) traditional segregation methods are confounded byphenocopies and are evenmore difficult to employwith genesthat have reduced penetrance These confounders indicatethat these genes require large numbers of segregation eventsto provide meaningful results Consideration should also begiven to gene-specific factors such as frequency of germlineand somatic de novo alterations additional tests in tumorssuch as loss of heterozygosity studies variation in nonsense-mediated decay and alternate splicing For example in genessuch asTP53 and PTEN germline de novo variants are knownto be a relatively common cause of disease [11 12] Howeverwith breast cancer genes such as ATM CHEK2 and PALB2the de novo rate is unknown This is confounded by the factthat breast cancer is a common disease and one cannot inferif the de novo event in these genes directly correlates withdisease or occurred by chance In addition although somaticde novo data is available for some genes [13] its incorporationinto germline variant analysis has yet to be standardized andwill need to be performed on a gene-by-gene basis
Consortia such as the Evidence Based Network forthe Interpretation of Germline Mutant Alleles (ENIGMA)have demonstrated the power of a collaborative approachto variant assessment and have made great strides in thereclassification of VUS in breast cancer genes as pathogenicor benign However even these groups are limited by the rateat which data is accumulated Open-access databases suchas ClinVar and the Leiden Open (source) Variant Database(LOVD) can be useful in identifying additional cases orpublications related to a variant These databases have alsohelped standardize the interpretation of variants betweenlaboratories by identifying discrepancies in classificationsCollaborative efforts by clinical laboratories includingAmbryGenetics GeneDx University of Chicago and Laboratoryfor Molecular Medicine have resulted in the sharing ofinternal data consisting of segregation and cooccurrenceswith mutations in the same gene or other genes and de novo
observations have led to the resolution of 78 of clinicallyactionable differences (VUS versus VLPmutation) and 92of VUS versus likely benignbenign differences (internaldata) Despite these efforts one of the challenges faced bymolecular laboratories and clinicians is that many geneticvariants are very rare and do not have enough published datato be classified beyondVUSWe present here our laboratoryrsquosintegrated approach to variant assessment and review toolsused to assess the impact of variants on protein function
2 Integrated Approach to Variant Assessment
Ambry Genetics has developed and implemented an inte-grated approach to variant assessment (Table 1) that encom-passes a five-tier variant classification algorithm similar tothose presented by ACMG and IARC Although the founda-tion of Ambry Geneticsrsquo classification algorithm is based onthe ACMG guidelines we have adopted stringent thresholdssimilar to those proposed by the IARC where ldquolikelyrdquo refersto a gt95 confidence of a variant being disease-causing orbenign [9] In this algorithm both pathogenic and likelypathogenic variants are interpreted as clinically actionablewith recommendations for medical management and familymember testing
Ambry Geneticsrsquo algorithm incorporates multiple lines ofevidence aimed at assessing both the impact of the variant onthe protein and the pathogenicity of the variant in relationto a disease phenotype (Figure 1 and Table 2) These lines ofevidence are weighted as stand-alone (categories A and F)strong (categories B and D) or supportive (categories C andE) and when combined as described in Table 1 they can leadto a classification of likely benign benign likely pathogenicor pathogenic When the evidence is limited or conflictingthe variants remain classified as VUS Lines of evidence suchas its location structure-function and functional and RNAstudies reflect the functional impact on themRNAor proteinEvolutionary conservation in silicomodels such as Polyphenand SIFT and general population frequency reflect fitnessthat is reproductive success and survival as measured by alack of allelic diversity The observed phenotype in variantcarriers and the cosegregation of the variant with disease andthe cooccurrence with other pathogenic variants reflect thepathogenicity of the variant (Figure 1) Some of this evidenceis readily available via databases such as allele frequencydata in the Exome Aggregation Consortium (ExAC) or thedata in published literature [14] However published literaturegenerally contains data for common variants and the datasupporting pathogenicity for rare variants is scarce andfrequently only available internally
For most genes on breast cancer panels computationaldata from in silico models evolutionary conservation andprotein structural analysis are readily available Populationfrequency data has been accumulating at a fast pace dueto major contributions from 1000 Genomes NHLBI ExomeSequencing Project (ESP) and ExAC These data have hada significant impact on the identification of benign variantsat high frequencies that are too frequent to be pathogenicbased on disease incidence alone particularly for historicallyunderstudied ethnic groups For breast cancer genes this
International Journal of Breast Cancer 3
Table1Classifi
catio
nschemefor
high
penetrance
autosomaldo
minantb
reastcancerg
enes
Class
Classifi
catio
nCa
tegory
Criteria
5Pathogenic
A1n
eeded
(stand
-alone)
(i)Alteratio
nsresulting
inprem
aturetruncation(egreadingfram
eshift
nonsense)
(ii)O
ther
ACMG-definedmutations
(ie
initiationcodo
nor
grossd
eletion)
(iii)Strong
segregationwith
disease(LO
Dgt3=gt10
meioses)
5Pathogenic
B4needed
(stro
ng)
(i)Con
firmed
denovo
alteratio
nin
thes
ettin
gof
anew
disease(approp
riateph
enotype)in
thefam
ily(ii)S
ignificantd
iseasea
ssociatio
nin
approp
riatelysiz
edcase-con
trolstudy(ie
s)(iii)Be
ingdetected
inindividu
alssatisfying
establish
eddiagno
sticc
riteriaforc
lassicdiseasew
ithou
tacle
armutation
(iv)L
astn
ucleotideo
fexon
(v)G
oodsegregationwith
disease(LO
D15
ndash3=5ndash9meioses)
(vi)Deficientp
rotein
functio
nin
approp
riatefunctio
nalassay(s)
(vii)
Functio
nally
valid
ated
splicingmutation
(viii)W
ell-characterized
mutationatthes
amep
osition
(ix)O
ther
stron
gdatasupp
ortin
gpathogeniccla
ssificatio
n(egstr
uctural)
4Likelypathogenic
1needed
(i)Alteratio
nsatthec
anon
icaldo
noracceptor
sites
(plusmn12)
with
outano
ther
stron
g(B-le
vel)evidence
supp
ortin
gpathogenicity
4Likelypathogenic
C4needed
(sup
portive)
(i)Ra
rityin
generalp
opulationdatabases(db
SNPES
P1000
Genom
esE
xAC)
(ii)Insilico
mod
elsinagreem
ent(deleterio
us)a
ndorc
ompletely
conservedpo
sitionin
approp
riatespecies
(iii)Mod
erates
egregatio
nwith
disease(atleast3
inform
ativem
eioses)for
rare
diseases
(iv)O
ther
datasupp
ortin
gpathogeniccla
ssificatio
n(egstr
uctural)
3of
B2of
Bandatleast1
ofC
1ofB
andatleast3
ofC
3VUS
Insufficiento
rcon
flictingevidence
Grossdu
plications
with
outstro
ngevidence
forp
atho
genico
rbenign
3Likelybenign
D1n
eeded
(stro
ng)
(i)Intro
nica
lteratio
nwith
nosplicingim
pactby
RT-PCR
analysisor
anothersplicingassay
(ii)O
ther
stron
gdatasupp
ortin
gbenign
classificatio
n
3Likelybenign
E2needed
(sup
portive)
(i)Coo
ccurrences
with
mutations
inthes
ameg
ene(ph
aseu
nkno
wn)
(ii)C
ooccurrences
with
mutations
inotherh
ighpenetra
ntgenesthatclearlyexplainap
roband
rsquosph
enotype
(iii)Subp
opulationfre
quency
insupp
orto
fbenigncla
ssificatio
n(iv
)Intactp
rotein
functio
nob
served
inapprop
riatefunctio
nalassay(s)
(v)Insilico
mod
elsinagreem
ent(benign
)(vi)Not
segregatingwith
diseaseinfamily
study
(genes
with
incompletep
enetrance)
(vii)
Nodiseasea
ssociatio
nin
smallcase-controlstudy
(viii)O
ther
datasupp
ortin
gbenign
classificatio
n
1Be
nign
F1n
eeded
(stand
-alone)
(i)Generalpo
pulationor
subp
opulationfre
quency
istoohigh
tobe
apatho
genicm
utationbasedon
diseasesynd
romep
revalencea
ndpenetrance
(ii)N
otsegregatingwith
diseaseinfamily
study
(genes
with
completep
enetrance)
(iii)Internalfre
quency
istoohigh
tobe
apatho
genicm
utationbasedon
diseasesynd
romep
revalencea
ndpenetra
nce
(iv)B
eing
seen
intra
nswith
amutationor
inho
mozygou
sstatein
individu
alsw
ithou
tsevered
iseasefor
thatgene
(v)N
odiseasea
ssociatio
nin
approp
riatelysiz
edcase-con
trolstudy(ie
s)1o
fDandatleast2
ofE
2or
moreo
fDgt3of
Ewo
confl
ictin
gdata
gt4of
Ewcon
flictingdata
Thev
ariant
classificatio
nschemeisn
otintend
edforthe
interpretatio
nof
alteratio
nscomplicated
byepigeneticfactorsincluding
genetic
mod
ifiersmultifactoria
ldise
aseor
low-risk
diseasea
ssociatio
nallelesa
ndmay
belim
itedin
theinterpretationof
alteratio
nsconfou
nded
byincompletep
enetrancevaria
blee
xpressivitypheno
copiesand
triallelic
oroligogenicinheritance
4 International Journal of Breast Cancer
Variantclassification
Fitness
Patho
genicity
Function
In silicoassessment
Polyphen SIFTmutationAssessor
Provean ESE fruitfly
Evolutionaryconservation
UCSC genome browserNCBI
Populationfrequency data
1000 Genomes ESPExAC
Case-controlstudies
Published literatureinternal data
Phenotype
Published literatureinternal data
Cooccurrenceamp segregation
Published literatureinternal data
Structural function
Uniprot wwPDBinternal dataRNA studies
Published literatureinternal data
Functional assays
Published literatureinternal data
Figure 1 An integrated approach for variant classification Lines of evidence such as structural function RNA studies and functional studiesassess the functional impact on the mRNA and protein Cooccurrence segregation case-control studies and the observed phenotype invariant carriers reflect the pathogenicity of a variant Population frequency in silico models and evolutionary conservation assess fitness ofthe amino acid or nucleotide position
threshold has been conservatively set at an allele frequencyof 1 in large population cohorts if used as a stand-alone lineof evidence supporting benign classification (Table 1 categoryF) Careful consideration of population cohort size is neededto attain a high confidence (lower 95 CI is above 1 with119901 value lt005) that the frequency is above 1 For examplewith a cohort of 60000 alleles an allele frequency of 108is sufficient (lower 95 CI = 101 119901 = 00244) whereas fora cohort of 1000 alleles an allele frequency of 170 (lower95 CI = 115 119901 = 0013) is needed to be 95 confidentthe allele frequency is above 1
Although patient phenotype cooccurrence and cosegre-gation data can be found in the published literature manylaboratories also curate internal data for use in variant classi-fication A patientrsquos clinical and family history can be difficultto use as a line of evidence in a clinical laboratory settingdue to ascertainment bias However when a variant in a geneassociated with a rare disorder (less than 12000) is identifiedin multiple individuals meeting classic clinical criteria andnever in large control populations or population cohortsthese data can be used as evidence towards pathogenicityThis is most informative in patients who have undergonegenetic testing on large multigene panel tests in which allthe known genes associated with a disorder have been ruledout However when defining classic clinical criteria we usevery strict guidelines and exclude common diseases such as
breast cancer For example when assessing a TP53 variantthe phenotype is considered strong if the patientmeets classicLi-Fraumeni syndrome criteria a proband with sarcomadiagnosed before 45 years a first-degree relative with anycancer before 45 years and a second-degree relative with anycancer before age 45 years or a sarcoma at any age [15] Forcommon diseases and moderate penetrance genes Bayesiananalyses that require larger phenotype data sets are used [16]Historically in vitro studies were predominantly found in thepublished literature However due to the rapid accumulationof rare variants clinical laboratories such as Ambry Geneticsare implementing validated internal functional studies suchas splicing and homology-directed DNA break repair (HDR)assays that can be incorporated into variant classificationalgorithms
3 Functional Lab
Many variants are classified as VUS because their functionalimpact either is poorly understood or has not yet beeninvestigated These variants include missense and splicingalterations in tumor suppressor genes that require loss offunction to manifest a disease [7] Clinical genomic labora-tories have traditionally relied on evidence from publishedliterature to establish the impact of a variant on gene expres-sion or protein function [7] There are several limitations
International Journal of Breast Cancer 5
DNA-NGS test
Detection of a VUS
Functional lab
Missense VUS Splicing VUS Gross duplication
Functional assays
For example homologousrecombination
RNA studies
For example CE cloning andsequence analysis
Breakpoint studies
Analysis of insertionbreakpoints
Figure 2 Workflow of a functional lab for the evaluation of VUS
Table 2 Experimental structures of genes linked to breast cancerlowast
Gene Length PDBs Coverage ()ATM 3056 0 00BARD1 777 5 421BRCA1 1863 27 176BRCA2 3418 2 16BRIP1 1249 3 19CDH1 882 12 262CHEK2 543 38 864MRE11A 708 1 581MUTYH 549 2 773NBN 754 0 00NF1 2839 6 221PALB2 1186 2 297PTEN 403 6 928RAD50 1312 0 00RAD51C 376 0 00RAD51D 328 1 253TP53 393 142 1000lowastGene lengths and coverage are tabulated from the Universal ProteinResource (Uniprot) [4] and the Research Collaboratory for StructuralBioinformatics (RCSB) [5] databases The list of genes is taken from theBreastNext panel
to this approach including publication bias difficulties withpromptly obtaining additional information about results andprotocols and lack of published evidence for a specificalteration One potential solution is for clinical genomiclaboratories to implement a ldquofunctional labrdquo that can generateassays with high sensitivity and specificity (gt99) and pro-vide unbiased molecular evidence to elucidate the functionalimpact of a VUS (Figure 2) As an example of a convincinglyvalidated assay Guidugli and colleagues determined the
sensitivity of their homology-directed DNA break repair(HDR) functional assay to be 100 (95 confidence interval(CI) 753ndash100) and the specificity to be 100 (95 CI815ndash100) [17]
31 RNA Studies for Splicing VUS While some splicing vari-ants such as canonical plusmn1 or 2 splice sites are often assumedto disrupt gene function by leading to the reduced expressionof the abnormal allele due to nonsense-mediated decay(NMD) [18] or abnormal protein truncations [19] compre-hensive evaluation of splicing alterations is essential for accu-rate clinical interpretation For canonical splice site plusmn1 and 2variants onemust also consider the possibility of an in-framedeletioninsertion which could retain the critical regions ofthe protein and hence lead to a mild neutral or gain-of-function effect In addition variants that are predicted toimpact splicing but that are not located at the canonical sites(plusmn1 and 2) require additional strong evidence (see details inSection 2) to be classified as pathogenic or benign [7] Bioin-formatics software has been developed to predict putativesplice sites [20] In general these in silico tools are more sen-sitive (sim90ndash100) than being specific (sim60ndash80) when pre-dicting the impact of a variant on splicing [21 22] Howeverby nature in silico tools can only provide supporting evidencewhich restricts their use [7] Consequently data from RNAsplicing assays designed to provide quantitative and qualita-tive characterization of transcripts are usually necessary toevaluate the pathogenicity of these variants Since publishedRNA data is not available for every variant clinical genomiclaboratories can more accurately classify splicing alterationsby implementing their own RNA protocols and assays toprovide accurate classification of splicing alterations
Reliability in which an assay yields the same results inrepeated trials is a key issue when implementing mRNAassays in a clinical functional lab for evaluation of VUS
6 International Journal of Breast Cancer
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
Figure 1 An integrated approach for variant classification Lines of evidence such as structural function RNA studies and functional studiesassess the functional impact on the mRNA and protein Cooccurrence segregation case-control studies and the observed phenotype invariant carriers reflect the pathogenicity of a variant Population frequency in silico models and evolutionary conservation assess fitness ofthe amino acid or nucleotide position
threshold has been conservatively set at an allele frequencyof 1 in large population cohorts if used as a stand-alone lineof evidence supporting benign classification (Table 1 categoryF) Careful consideration of population cohort size is neededto attain a high confidence (lower 95 CI is above 1 with119901 value lt005) that the frequency is above 1 For examplewith a cohort of 60000 alleles an allele frequency of 108is sufficient (lower 95 CI = 101 119901 = 00244) whereas fora cohort of 1000 alleles an allele frequency of 170 (lower95 CI = 115 119901 = 0013) is needed to be 95 confidentthe allele frequency is above 1
Although patient phenotype cooccurrence and cosegre-gation data can be found in the published literature manylaboratories also curate internal data for use in variant classi-fication A patientrsquos clinical and family history can be difficultto use as a line of evidence in a clinical laboratory settingdue to ascertainment bias However when a variant in a geneassociated with a rare disorder (less than 12000) is identifiedin multiple individuals meeting classic clinical criteria andnever in large control populations or population cohortsthese data can be used as evidence towards pathogenicityThis is most informative in patients who have undergonegenetic testing on large multigene panel tests in which allthe known genes associated with a disorder have been ruledout However when defining classic clinical criteria we usevery strict guidelines and exclude common diseases such as
breast cancer For example when assessing a TP53 variantthe phenotype is considered strong if the patientmeets classicLi-Fraumeni syndrome criteria a proband with sarcomadiagnosed before 45 years a first-degree relative with anycancer before 45 years and a second-degree relative with anycancer before age 45 years or a sarcoma at any age [15] Forcommon diseases and moderate penetrance genes Bayesiananalyses that require larger phenotype data sets are used [16]Historically in vitro studies were predominantly found in thepublished literature However due to the rapid accumulationof rare variants clinical laboratories such as Ambry Geneticsare implementing validated internal functional studies suchas splicing and homology-directed DNA break repair (HDR)assays that can be incorporated into variant classificationalgorithms
3 Functional Lab
Many variants are classified as VUS because their functionalimpact either is poorly understood or has not yet beeninvestigated These variants include missense and splicingalterations in tumor suppressor genes that require loss offunction to manifest a disease [7] Clinical genomic labora-tories have traditionally relied on evidence from publishedliterature to establish the impact of a variant on gene expres-sion or protein function [7] There are several limitations
International Journal of Breast Cancer 5
DNA-NGS test
Detection of a VUS
Functional lab
Missense VUS Splicing VUS Gross duplication
Functional assays
For example homologousrecombination
RNA studies
For example CE cloning andsequence analysis
Breakpoint studies
Analysis of insertionbreakpoints
Figure 2 Workflow of a functional lab for the evaluation of VUS
Table 2 Experimental structures of genes linked to breast cancerlowast
Gene Length PDBs Coverage ()ATM 3056 0 00BARD1 777 5 421BRCA1 1863 27 176BRCA2 3418 2 16BRIP1 1249 3 19CDH1 882 12 262CHEK2 543 38 864MRE11A 708 1 581MUTYH 549 2 773NBN 754 0 00NF1 2839 6 221PALB2 1186 2 297PTEN 403 6 928RAD50 1312 0 00RAD51C 376 0 00RAD51D 328 1 253TP53 393 142 1000lowastGene lengths and coverage are tabulated from the Universal ProteinResource (Uniprot) [4] and the Research Collaboratory for StructuralBioinformatics (RCSB) [5] databases The list of genes is taken from theBreastNext panel
to this approach including publication bias difficulties withpromptly obtaining additional information about results andprotocols and lack of published evidence for a specificalteration One potential solution is for clinical genomiclaboratories to implement a ldquofunctional labrdquo that can generateassays with high sensitivity and specificity (gt99) and pro-vide unbiased molecular evidence to elucidate the functionalimpact of a VUS (Figure 2) As an example of a convincinglyvalidated assay Guidugli and colleagues determined the
sensitivity of their homology-directed DNA break repair(HDR) functional assay to be 100 (95 confidence interval(CI) 753ndash100) and the specificity to be 100 (95 CI815ndash100) [17]
31 RNA Studies for Splicing VUS While some splicing vari-ants such as canonical plusmn1 or 2 splice sites are often assumedto disrupt gene function by leading to the reduced expressionof the abnormal allele due to nonsense-mediated decay(NMD) [18] or abnormal protein truncations [19] compre-hensive evaluation of splicing alterations is essential for accu-rate clinical interpretation For canonical splice site plusmn1 and 2variants onemust also consider the possibility of an in-framedeletioninsertion which could retain the critical regions ofthe protein and hence lead to a mild neutral or gain-of-function effect In addition variants that are predicted toimpact splicing but that are not located at the canonical sites(plusmn1 and 2) require additional strong evidence (see details inSection 2) to be classified as pathogenic or benign [7] Bioin-formatics software has been developed to predict putativesplice sites [20] In general these in silico tools are more sen-sitive (sim90ndash100) than being specific (sim60ndash80) when pre-dicting the impact of a variant on splicing [21 22] Howeverby nature in silico tools can only provide supporting evidencewhich restricts their use [7] Consequently data from RNAsplicing assays designed to provide quantitative and qualita-tive characterization of transcripts are usually necessary toevaluate the pathogenicity of these variants Since publishedRNA data is not available for every variant clinical genomiclaboratories can more accurately classify splicing alterationsby implementing their own RNA protocols and assays toprovide accurate classification of splicing alterations
Reliability in which an assay yields the same results inrepeated trials is a key issue when implementing mRNAassays in a clinical functional lab for evaluation of VUS
6 International Journal of Breast Cancer
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
Figure 1 An integrated approach for variant classification Lines of evidence such as structural function RNA studies and functional studiesassess the functional impact on the mRNA and protein Cooccurrence segregation case-control studies and the observed phenotype invariant carriers reflect the pathogenicity of a variant Population frequency in silico models and evolutionary conservation assess fitness ofthe amino acid or nucleotide position
threshold has been conservatively set at an allele frequencyof 1 in large population cohorts if used as a stand-alone lineof evidence supporting benign classification (Table 1 categoryF) Careful consideration of population cohort size is neededto attain a high confidence (lower 95 CI is above 1 with119901 value lt005) that the frequency is above 1 For examplewith a cohort of 60000 alleles an allele frequency of 108is sufficient (lower 95 CI = 101 119901 = 00244) whereas fora cohort of 1000 alleles an allele frequency of 170 (lower95 CI = 115 119901 = 0013) is needed to be 95 confidentthe allele frequency is above 1
Although patient phenotype cooccurrence and cosegre-gation data can be found in the published literature manylaboratories also curate internal data for use in variant classi-fication A patientrsquos clinical and family history can be difficultto use as a line of evidence in a clinical laboratory settingdue to ascertainment bias However when a variant in a geneassociated with a rare disorder (less than 12000) is identifiedin multiple individuals meeting classic clinical criteria andnever in large control populations or population cohortsthese data can be used as evidence towards pathogenicityThis is most informative in patients who have undergonegenetic testing on large multigene panel tests in which allthe known genes associated with a disorder have been ruledout However when defining classic clinical criteria we usevery strict guidelines and exclude common diseases such as
breast cancer For example when assessing a TP53 variantthe phenotype is considered strong if the patientmeets classicLi-Fraumeni syndrome criteria a proband with sarcomadiagnosed before 45 years a first-degree relative with anycancer before 45 years and a second-degree relative with anycancer before age 45 years or a sarcoma at any age [15] Forcommon diseases and moderate penetrance genes Bayesiananalyses that require larger phenotype data sets are used [16]Historically in vitro studies were predominantly found in thepublished literature However due to the rapid accumulationof rare variants clinical laboratories such as Ambry Geneticsare implementing validated internal functional studies suchas splicing and homology-directed DNA break repair (HDR)assays that can be incorporated into variant classificationalgorithms
3 Functional Lab
Many variants are classified as VUS because their functionalimpact either is poorly understood or has not yet beeninvestigated These variants include missense and splicingalterations in tumor suppressor genes that require loss offunction to manifest a disease [7] Clinical genomic labora-tories have traditionally relied on evidence from publishedliterature to establish the impact of a variant on gene expres-sion or protein function [7] There are several limitations
International Journal of Breast Cancer 5
DNA-NGS test
Detection of a VUS
Functional lab
Missense VUS Splicing VUS Gross duplication
Functional assays
For example homologousrecombination
RNA studies
For example CE cloning andsequence analysis
Breakpoint studies
Analysis of insertionbreakpoints
Figure 2 Workflow of a functional lab for the evaluation of VUS
Table 2 Experimental structures of genes linked to breast cancerlowast
Gene Length PDBs Coverage ()ATM 3056 0 00BARD1 777 5 421BRCA1 1863 27 176BRCA2 3418 2 16BRIP1 1249 3 19CDH1 882 12 262CHEK2 543 38 864MRE11A 708 1 581MUTYH 549 2 773NBN 754 0 00NF1 2839 6 221PALB2 1186 2 297PTEN 403 6 928RAD50 1312 0 00RAD51C 376 0 00RAD51D 328 1 253TP53 393 142 1000lowastGene lengths and coverage are tabulated from the Universal ProteinResource (Uniprot) [4] and the Research Collaboratory for StructuralBioinformatics (RCSB) [5] databases The list of genes is taken from theBreastNext panel
to this approach including publication bias difficulties withpromptly obtaining additional information about results andprotocols and lack of published evidence for a specificalteration One potential solution is for clinical genomiclaboratories to implement a ldquofunctional labrdquo that can generateassays with high sensitivity and specificity (gt99) and pro-vide unbiased molecular evidence to elucidate the functionalimpact of a VUS (Figure 2) As an example of a convincinglyvalidated assay Guidugli and colleagues determined the
sensitivity of their homology-directed DNA break repair(HDR) functional assay to be 100 (95 confidence interval(CI) 753ndash100) and the specificity to be 100 (95 CI815ndash100) [17]
31 RNA Studies for Splicing VUS While some splicing vari-ants such as canonical plusmn1 or 2 splice sites are often assumedto disrupt gene function by leading to the reduced expressionof the abnormal allele due to nonsense-mediated decay(NMD) [18] or abnormal protein truncations [19] compre-hensive evaluation of splicing alterations is essential for accu-rate clinical interpretation For canonical splice site plusmn1 and 2variants onemust also consider the possibility of an in-framedeletioninsertion which could retain the critical regions ofthe protein and hence lead to a mild neutral or gain-of-function effect In addition variants that are predicted toimpact splicing but that are not located at the canonical sites(plusmn1 and 2) require additional strong evidence (see details inSection 2) to be classified as pathogenic or benign [7] Bioin-formatics software has been developed to predict putativesplice sites [20] In general these in silico tools are more sen-sitive (sim90ndash100) than being specific (sim60ndash80) when pre-dicting the impact of a variant on splicing [21 22] Howeverby nature in silico tools can only provide supporting evidencewhich restricts their use [7] Consequently data from RNAsplicing assays designed to provide quantitative and qualita-tive characterization of transcripts are usually necessary toevaluate the pathogenicity of these variants Since publishedRNA data is not available for every variant clinical genomiclaboratories can more accurately classify splicing alterationsby implementing their own RNA protocols and assays toprovide accurate classification of splicing alterations
Reliability in which an assay yields the same results inrepeated trials is a key issue when implementing mRNAassays in a clinical functional lab for evaluation of VUS
6 International Journal of Breast Cancer
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
Figure 2 Workflow of a functional lab for the evaluation of VUS
Table 2 Experimental structures of genes linked to breast cancerlowast
Gene Length PDBs Coverage ()ATM 3056 0 00BARD1 777 5 421BRCA1 1863 27 176BRCA2 3418 2 16BRIP1 1249 3 19CDH1 882 12 262CHEK2 543 38 864MRE11A 708 1 581MUTYH 549 2 773NBN 754 0 00NF1 2839 6 221PALB2 1186 2 297PTEN 403 6 928RAD50 1312 0 00RAD51C 376 0 00RAD51D 328 1 253TP53 393 142 1000lowastGene lengths and coverage are tabulated from the Universal ProteinResource (Uniprot) [4] and the Research Collaboratory for StructuralBioinformatics (RCSB) [5] databases The list of genes is taken from theBreastNext panel
to this approach including publication bias difficulties withpromptly obtaining additional information about results andprotocols and lack of published evidence for a specificalteration One potential solution is for clinical genomiclaboratories to implement a ldquofunctional labrdquo that can generateassays with high sensitivity and specificity (gt99) and pro-vide unbiased molecular evidence to elucidate the functionalimpact of a VUS (Figure 2) As an example of a convincinglyvalidated assay Guidugli and colleagues determined the
sensitivity of their homology-directed DNA break repair(HDR) functional assay to be 100 (95 confidence interval(CI) 753ndash100) and the specificity to be 100 (95 CI815ndash100) [17]
31 RNA Studies for Splicing VUS While some splicing vari-ants such as canonical plusmn1 or 2 splice sites are often assumedto disrupt gene function by leading to the reduced expressionof the abnormal allele due to nonsense-mediated decay(NMD) [18] or abnormal protein truncations [19] compre-hensive evaluation of splicing alterations is essential for accu-rate clinical interpretation For canonical splice site plusmn1 and 2variants onemust also consider the possibility of an in-framedeletioninsertion which could retain the critical regions ofthe protein and hence lead to a mild neutral or gain-of-function effect In addition variants that are predicted toimpact splicing but that are not located at the canonical sites(plusmn1 and 2) require additional strong evidence (see details inSection 2) to be classified as pathogenic or benign [7] Bioin-formatics software has been developed to predict putativesplice sites [20] In general these in silico tools are more sen-sitive (sim90ndash100) than being specific (sim60ndash80) when pre-dicting the impact of a variant on splicing [21 22] Howeverby nature in silico tools can only provide supporting evidencewhich restricts their use [7] Consequently data from RNAsplicing assays designed to provide quantitative and qualita-tive characterization of transcripts are usually necessary toevaluate the pathogenicity of these variants Since publishedRNA data is not available for every variant clinical genomiclaboratories can more accurately classify splicing alterationsby implementing their own RNA protocols and assays toprovide accurate classification of splicing alterations
Reliability in which an assay yields the same results inrepeated trials is a key issue when implementing mRNAassays in a clinical functional lab for evaluation of VUS
6 International Journal of Breast Cancer
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
To improve reliability the ENIGMA consortium conducteda multicenter investigation to compare mRNA splicing assayprotocols used by its members [23] The consortium pro-vided several recommendations for best practices in clinicaltesting of splicing alterations including the standardizationof protocols and the use of analytically sensitive detectionmethods [23] Of the detection methods evaluated capillaryelectrophoresis (CE) was shown to yield the highest analyticsensitivity However a major limitation of CE is its inabilityto harvest and subsequently perform sequence analysis of theRT-PCR product In order to perform sequence analysis andfull characterization of alternatively spliced transcripts theconsortium concluded that cloning single PCR products intoa vector system is a useful alternative for isolating single tran-scripts for sequencing which improves sensitivity over bandexcision and sequencing alone Even in cases that appearstraightforward the consortium recommends using in vivoin vitro and clinical analysis to predict with 99 likelihoodthat a variant is benign or pathogenic [23] For examplealthough most canonical splice site variants are considered apriori to be at least likely pathogenic the presence of naturallyoccurring alternative splicing that mimics a pathogenicalteration and results in a similar impact on splicing (egexon skipping) needs to be carefully evaluated as it mayresult in diminished pathogenicity Care must be taken todetermine whether a transcript is present in normal controlsAs the functional lab obtains more data on each gene a moreaccurate picture of splicing patterns will emerge therebyleading to improved classification of splice site variations
32 Functional Assays forMissenseVUS Missense alterationswith no impact on splicing can be evaluated by utilizing wetlab assays or experimental structure data While functionalstudies can be a powerful tool in support of pathogenicity notall functional studies accurately predict impacts on gene orprotein function For this reason ACMGAMP provides rec-ommendations for assessing the validity of functional assaysin order to confirm that the functional assay accurately mea-sures a function that leads to disease [7] One must considerhow closely the functional assay reflects the biological envi-ronment This is important when deciding whether to testpatient samples or to perform in vitro assays It is important toconsider the known biological functions of the protein whilealso examining whether those functions actually contributeto tumorigenicity For example many functional assays havebeen developed to interrogate BRCA1VUS [24] Some assaysfocus on the known DNA repair functions of BRCA1 suchas the HDR assay [25 26] and the radiation resistance assay[27] Others examine BRCA1 localization [28 29] and theability of cells with BRCA1 variants to generate Rad51 foci[30 31] in the presence of DNA damage as surrogates forBRCA1 function Additional assays focus on one functionalcomponent of BRCA1 instead of the full protein including thetranscription activation assay which employs the C-terminalBRCT domains and the ubiquitin ligase assay which utilizesthe N-terminal region [32ndash35] These two assays are limitedby their inability to account for effects of the entire protein
and others have noted that certain variants that lost ubiquitinligase activity were not classified as pathogenic by geneticstudies [36 37] Similarly protein or peptide binding assaysmay resolve the ability of a variant to bind to a protein targetin vitro but these data should be incorporated into a mul-tifactorial model that takes into account other functional invivo data [38 39] In addition validation data that assess theanalytical performance of the assay and account for specimenintegrity are important factors to consider when implement-ing functional assays in a functional clinical genomic labora-tory and in using these results in classification of variants [7]
To investigate the effect of missense variants on BRCA1function Lu et al tested 68missense variants using an in vitroHDR assay [26] The analysis showed that the HDR defectiveor partial defective missense variants from the BRCT domainare positioned either in the center of the structure or onthe surface responsible for protein-protein interactions whilethe HDR-WT variants from the BRCT domain were surfaceexposed or partially surface exposed variants [26] Thishighlights the complexity of interpreting missense germlinevariants indicating that an integrated approach by compilingthe results of functional assays structure evaluation andanalysis of clinical parameters should identify the mostfunctionally and clinically relevant alterations
33 Analysis of Insertion Breakpoints for Gross DuplicationsMost gross deletions in high-risk cancer genes larger than 3sim5 megabases fall within microarray reporting guidelines andare reported as deleterious [38 39] however without break-point information gross duplications are mostly reportedas VUS While array comparative genomic hybridization(aCGH) is a method used in cancer research for the detec-tion of gross chromosomal aberrations in cancer genes itcannot accurately determine the exact genomic breakpointsof the amplification [40ndash43] To map the exact insertionbreakpoints paired-end high throughput sequencing can beused Gross genomic amplifications may occur as a tandemduplication within the cancer gene itself resulting in a novelfunction or as a nontandem duplication inserted in a novellocation of the genome Therefore identifying the exactbreakpoints of tandem duplications in high-risk cancer genescan lead to VUS being reclassified as likely pathogenic orlikely benign
To identify the exact breakpoints of tandem duplica-tions Ambry Genetics is currently utilizing the paired-endsequencing method to further characterize gross duplica-tions Probe sets are designed to capture the target regionswith the suggested breakpoints identified by aCGHCapturedDNA is then sequenced by paired-end massively parallelsequencing and mapped to the human genome The AmbryGenetics pipeline identifies read pairs that are in the wrongorientation indicating a tandem duplication (Figure 3(a))Clusters of read pairs with soft clipping that span breakpointscan indicate rearrangement breakpoints down to the exactcoordinates (Figure 3(b)) As an example an exon 11 duplica-tion in BRCA1 previously classified as VUS can be reclassifiedas likely pathogenic if the breakpoint is identified to cause aframeshift in the gene (Figure 3)
International Journal of Breast Cancer 7
Tandem duplication
Reference
BRCA1 gene
Paired sequence read aligned to target gene
Exon 11
(a)
BRCA1 Exon 11
(b)
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
Figure 3 Identification of tandem duplication insertion breakpoints spanning BRCA1 exon 11 using paired-end sequencing (a) Mappedread pairs in the wrong orientation indicate a tandem duplication (b) Ambryrsquos breakpoint detection tools can identify clusters of read pairswith soft clipping which indicate rearrangement breakpoints
4 Computational Structural Analysis
Computational structural algorithms offer a unique solutionfor assessing a variantrsquos impact on protein function in thatthey are faster than experimental studies and often use datafrommany scientific disciplines [44] However the quality ofthe information provided by computational analyses variesdepending on the information source For instance primarysequence analyses using evolutionary tools can identify thelikely impact of a variant By comparing an altered humansequence to proteins with a similar primary sequence orrelated structural shape the fitness of the variant can bepredicted based on the variability of that position and otheraspects such as the chemical similarity of the wild type andvariant amino acids Ambry Genetics relies onmultiple toolsincluding the ldquoSorting Intolerant from Tolerantrdquo (SIFT) andPolyphen2 programs [40] We use the consensus of twoprograms usually SIFT and Polyphen2 where applicable andconsider only concordant results as a line of evidence Ifonly one program is applicable such as Provean [45] withindels we incorporate predictions from the single programwith conservative thresholds determined by analysis of ourinternal data Alternatively analyses of the secondary and ter-tiary structures of the protein increase the reliability of inter-pretation The most reliable computational algorithms focuson biophysics methods which are more oriented towardsdirect simulation of the physical processes occurring in aprotein [46ndash48] In many regards computational methodsare the most diverse in the range of properties that they canquantify however they come at the expense of computationalrequirements and speed with which accurate properties canbe derived One of the most common and easily identified
sources of disruption induced by a variant is the influenceon protein stability [46 47 49ndash51] Protein stability can beaffected in multiple ways such as misfolding or unfolding ofthe protein structure which commonly results in either lossof function or premature degradation and haploinsufficiencyAs an example protein stability has been used by Karchin etal to generate a predictive tool for the likelihood of the effectof an alteration in the breast cancer gene BRCA2 [52] Thereare other significant ways that variants exert their pathogeniceffect which can be described through structure For instancea variantmaynot significantly affect the resting state structureof the protein but rather affect the movement of the proteinin the course of its function It may impact its bindingwith other target proteins or substrates or it may induceaggregation [47 48] Detailed understanding of biophysicalprinciples illuminated through structure is crucial to evaluateand interpret the impact of alterations
5 Tertiary and Quaternary Sequence ofBreast Cancer Genes
Theuse of biophysicalmethods to predict the impact of a vari-ant on a protein often requires the availability of structures forthe target gene or benefits significantly from it Among the 17genes represented in the BreastNext Cancer panel there are atotal of 247 experimentally derived structures tabulated pergene in the PDBs (Protein DataBank files) column of Table 2using either Nuclear Magnetic Resonance Imaging (NMR)or X-ray crystallographic methods [5 53] The coveragedescribed above corresponds to the total range of residuescovered by all experimental measurements divided by thetotal length of the proteinWhile there are notable exceptions
8 International Journal of Breast Cancer
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
where no experimental structures have been determined themajority of the genes have been partially and in the caseof TP53 completely elucidated experimentally The structureof TP53 is highly ordered throughout the protein allowingfor complete measurement of one low-energy form howeversome proteins in this set such as BRCA1 and BRCA2 are com-posed of regions which have no characteristic fixed structureThe ordered regions within the structured protein such asin the N (Really Interesting New Gene ie RING domain)and C terminus (BRCA1 C Terminus ie BRCT repeats)of BRCA1 offer higher quality means to define domainboundariesThese can be analyzed as a folded functional unitrather than through conservation techniques that are used inthe Protein Families (Pfam) database [54] or using meta pre-dictors such as InterPro [55] The structural coverage for thegenes in Table 2 does not take into account that long stretchesof some proteins have little intrinsic globular structure sothe numbers can be seen as a very conservative estimate ofthe range of available residues covered In addition thereremain someproteins such asATM orNBN where no or low-resolution structures have been experimentally measured [56] For these systems structural analysis incorporates the useof homology models built on the structures of known relatedproteins This significantly increases the effective range ofstructural coverage and the insights available
Disruption in the folding of a domain in a protein bya missense pathogenic variant is well known to result in aloss of functionThe clinically observed alteration c5509TgtG(pTrp1837Gly) (ClinVar SCV000077040) represents a casewhere structural features explain the disruption of the BRCTrepeat region in BRCA1 The C-terminal portion of BRCA1contains a pair of BRCT repeat domains BRCT1 and BRCT2which are described in atomic detail including the arrange-ment of amino acids that make up these domains by 26different crystal structures [56ndash59] The side chain of aminoacid Trp 1837 (W1837 magenta stick) is buried in the coreof the BRCT2 domain surrounded by hydrophobic aminoacids (green sticks) while the backbone participates in a helixinvolved in binding BACH1 (Figure 4) [56] The alterationW1837 to G1837 (W1837G) would result in the loss of thelarge stabilizing hydrophobic side chain and is anticipatedto be very destabilizing The instability introduced by thisalteration has been quantitatively calculated by computa-tional folding algorithms which indicate it to be very desta-bilizing [5] Indeed E coli expressed with in vitro mutantsof pW1837G produce an unfolded protein that was presentonly in inclusion bodies which could not be refolded [60]In another set of biochemical and cell-based transcriptionalexperiments this alteration resulted in compromised prote-olysis and phosphopeptide-binding [39 58 59 61] Togetherthese functional data support the qualitative and quantitativestructural observation that the variant would create a veryunfavorable cavity within this domain thereby disruptingfolding and protein functionThis example demonstrates howdetailed structural analysis on publically available data canfacilitate the understanding and interpretation of alterationson the function of a protein and can be supported by bothcomputational and experimental observations
BACH1
pW1837G
BRCA1-BRCT2
Figure 4 The structure of BRCA1 pTrp1837 (shown in magentawith sticks) in the BRCA-BRCT domain (PDB 1T15 [6]) Nearbyhydrophobic amino acids sidechains from residue 1837 are shown assticks Bound BACH1 peptide is shown as teal stick
6 Conclusion
Although cancer genetic testing has traditionally been limitedto highly penetrant and well-characterized susceptibilitygenes the application of multigene panels using massivelyparallel sequencing is steadily becoming more common ingenetic cancer risk assessment due to reduced costs andincreased efficiency Multigene panels in turn tend to resultin the identification of more variants per individual theclinical significance of which needs to be assessed usingmultiple lines of weighted evidenceWe present an integratedapproach for assessing variants observed on hereditary breastcancer panels and believe that this improves the clinicalmanagement of patients with personal and family historiesof breast cancer due to more accurate variant classificationComprehensive variant assessment programs that integratemultiple lines of evidence aimed at assessing a variantrsquosimpact on protein function fitness and pathogenicity facili-tate high-quality and efficient variant classification providingincreased benefit and reliability for patients
Competing Interests
Ambry Genetics Corp is a CLIA approved clinical geneticstesting laboratory The authors of this review are all full timepaid employees of Ambry Genetics
References
[1] H LaDuca A J Stuenkel J S Dolinsky et al ldquoUtilization ofmultigene panels in hereditary cancer predisposition testinganalysis of more than 2000 patientsrdquo Genetics in Medicine vol16 no 11 pp 830ndash837 2014
International Journal of Breast Cancer 9
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
[2] M B Daly J E Axilbund S Buys et al ldquoGeneticfamilial high-risk assessment breast and ovarianrdquo The National Comprehen-sive Cancer Network vol 8 pp 562ndash594 2010
[3] ldquoNational Comprehensive Cancer Network GeneticFamilialHigh-Risk Assessment Breast and Ovarian (Version 22016)rdquo2016 httpswwwnccnorgprofessionalsphysician glspdfgenetics screeningpdf
[4] M Magrane and UniProt Consortium ldquoUniProt knowledge-base a hub of integrated protein datardquo Database vol 2011Article ID bar009 2011
[5] H M Berman J Westbrook Z Feng et al ldquoThe protein databankrdquo Nucleic Acids Research vol 28 no 1 pp 235ndash242 2000
[6] W C Lau Y Li Z Liu Y Gao Q Zhang and M S HuenldquoStructure of the human dimeric ATM kinaserdquo Cell Cycle vol15 no 8 pp 1117ndash1124 2016
[7] S Richards N Aziz S Bale et al ldquoStandards and guidelinesfor the interpretation of sequence variants a joint consensusrecommendation of the American College of Medical Geneticsand Genomics and the Association for Molecular PathologyrdquoGenetics in Medicine vol 17 no 5 pp 405ndash424 2015
[8] C S Richards S Bale D B Bellissimo et al ldquoACMG recom-mendations for standards for interpretation and reporting ofsequence variations revisions 2007rdquo Genetics in Medicine vol10 no 4 pp 294ndash300 2008
[9] S V Tavtigian M S Greenblatt D E Goldgar and P BoffettaldquoAssessing pathogenicity overview of results from the IARCunclassified genetic variants working grouprdquoHumanMutationvol 29 no 11 pp 1261ndash1264 2008
[10] K N Maxwell S N Hart J Vijai et al ldquoEvaluation of ACMG-guideline-based variant classification of cancer susceptibilityand non-cancer-associated genes in families affected by breastcancerrdquoTheAmerican Journal of Human Genetics vol 98 no 5pp 801ndash817 2016
[11] K D Gonzalez C H Buzin K A Noltner et al ldquoHighfrequency of de novo mutations in Li-Fraumeni syndromerdquoJournal of Medical Genetics vol 46 no 10 pp 689ndash693 2009
[12] J Mester and C Eng ldquoEstimate of de novo mutation frequencyin probands with PTENhamartoma tumor syndromerdquoGeneticsin Medicine vol 14 no 9 pp 819ndash822 2012
[13] D R Walker J P Bond R E Tarone et al ldquoEvolutionaryconservation and somatic mutation hotspot maps of p53correlation with p53 protein structural and functional featuresrdquoOncogene vol 18 no 1 pp 211ndash218 1999
[14] M Lek K J Karczewski E V Minikel et al ldquoAnalysis ofprotein-coding genetic variation in 60706 humansrdquo Naturevol 536 no 7616 pp 285ndash291 2016
[15] F P Li J F Fraumeni Jr J J Mulvihill et al ldquoA cancer familysyndrome in twenty-four kindredsrdquo Cancer Research vol 48no 18 pp 5358ndash5362 1988
[16] D Pruss B Morris E Hughes et al ldquoDevelopment andvalidation of a new algorithm for the reclassification of geneticvariants identified in the BRCA1 and BRCA2 genesrdquo BreastCancer Research and Treatment vol 147 no 1 pp 119ndash132 2014
[17] L Guidugli V S Pankratz N Singh et al ldquoA classificationmodel for BRCA2 DNA binding domain missense variantsbased on homology-directed repair activityrdquo Cancer Researchvol 73 no 1 pp 265ndash275 2013
[18] R Karam J Wengrod L B Gardner and M F WilkinsonldquoRegulation of nonsense-mediated mRNA decay implicationsfor physiology and diseaserdquo Biochimica et Biophysica ActamdashGene Regulatory Mechanisms vol 1829 no 6-7 pp 624ndash6332013
[19] O Anczukow M D Ware M Buisson et al ldquoDoes thenonsense-mediated mRNA decay mechanism prevent the syn-thesis of truncated BRCA1 CHK2 and p53 proteinsrdquo HumanMutation vol 29 no 1 pp 65ndash73 2008
[20] X Jian E Boerwinkle and X Liu ldquoIn silico tools for splicingdefect prediction a survey from the viewpoint of end usersrdquoGenetics in Medicine vol 16 no 7 pp 497ndash503 2014
[21] C Houdayer V Caux-Moncoutier S Krieger et al ldquoGuidelinesfor splicing analysis inmolecular diagnosis derived from a set of327 combined in silicoin vitro studies on BRCA1 and BRCA2variantsrdquo Human Mutation vol 33 no 8 pp 1228ndash1238 2012
[22] M P G Vreeswijk J N Kraan H M van der Klift et alldquoIntronic variants in BRCA1 and BRCA2 that affect RNA splic-ing can be reliably selected by splice-site prediction programsrdquoHuman Mutation vol 30 no 1 pp 107ndash114 2009
[23] P J Whiley M de la Hoya M Thomassen et al ldquoComparisonof mRNA splicing assay protocols across multiple laboratoriesrecommendations for best practice in standardized clinicaltestingrdquo Clinical Chemistry vol 60 no 2 pp 341ndash352 2014
[24] G A Millot M A Carvalho S M Caputo et al ldquoA guidefor functional analysis of BRCA1 variants of uncertain signifi-cancerdquo Human Mutation vol 33 no 11 pp 1526ndash1537 2012
[25] D J R Ransburgh N Chiba C Ishioka A E Toland and JD Parvin ldquoIdentification of breast tumor mutations in BRCA1that abolish its function in homologous DNA recombinationrdquoCancer Research vol 70 no 3 pp 988ndash995 2010
[26] C Lu M Xie M C Wendl et al ldquoPatterns and functionalimplications of rare germline variants across 12 cancer typesrdquoNature Communications vol 6 Article ID 10086 2015
[27] R Scully S Ganesan K Vlasakova J Chen M Socolovskyand D M Livingston ldquoGenetic analysis of BRCA1 function in adefined tumor cell linerdquo Molecular Cell vol 4 no 6 pp 1093ndash1099 1999
[28] WWYAu andB RHenderson ldquoTheBRCA1RINGandBRCTdomains cooperate in targeting BRCA1 to ionizing radiation-induced nuclear focirdquo The Journal of Biological Chemistry vol280 no 8 pp 6993ndash7001 2005
[29] J A RodriguezWWYAu andB RHenderson ldquoCytoplasmicmislocalization of BRCA1 caused by cancer-associated muta-tions in theBRCTdomainrdquoExperimental Cell Research vol 293no 1 pp 14ndash21 2004
[30] R Drost P Bouwman S Rottenberg et al ldquoBRCA1 RINGfunction is essential for tumor suppression but dispensable fortherapy resistancerdquoCancer Cell vol 20 no 6 pp 797ndash809 2011
[31] K A TNaipal N S Verkaik N Ameziane et al ldquoFunctional Exvivo assay to select homologous recombination-deficient breasttumors for PARP inhibitor treatmentrdquoClinical Cancer Researchvol 20 no 18 pp 4816ndash4826 2014
[32] A N A Monteiro A August and H Hanafusa ldquoEvidencefor a transcriptional activation function of BRCA1 C-terminalregionrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 93 no 24 pp 13595ndash13599 1996
[33] J Vallon-Christersson C Cayanan K Haraldsson et al ldquoFunc-tional analysis of BRCA1 C-terminal missense mutations iden-tified in breast and ovarian cancer familiesrdquo Human MolecularGenetics vol 10 no 4 pp 353ndash360 2001
[34] C M Phelan V Dapic B Tice et al ldquoClassification of BRCA1missense variants of unknown clinical significancerdquo Journal ofMedical Genetics vol 42 no 2 pp 138ndash146 2005
[35] M A Carvalho S M Marsillac R Karchin et al ldquoDetermina-tion of cancer risk associated with germ line BRCA1 missense
10 International Journal of Breast Cancer
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006
variants by functional analysisrdquo Cancer Research vol 67 no 4pp 1494ndash1501 2007
[36] P S Brzovic J R Keeffe H Nishikawa et al ldquoBindingand recognition in the assembly of an active BRCA1BARD1ubiquitin-ligase complexrdquo Proceedings of the National Academyof Sciences of the United States of America vol 100 no 10 pp5646ndash5651 2003
[37] J R Morris L Pangon C Boutell T Katagiri N H Keep andE Solomon ldquoGenetic analysis of BRCA1 ubiquitin ligase activityand its relationship to breast cancer susceptibilityrdquo HumanMolecular Genetics vol 15 no 4 pp 599ndash606 2006
[38] S T South C Lee A N Lamb et al ldquoACMG standards andguidelines for constitutional cytogenomic microarray analysisincluding postnatal and prenatal applications revision 2013rdquoGenetics in Medicine vol 15 no 11 pp 901ndash909 2013
[39] M S Lee R Green S M Marsillac et al ldquoComprehensiveanalysis of missense variations in the BRCT domain of BRCA1by structural and functional assaysrdquo Cancer Research vol 70no 12 pp 4880ndash4890 2010
[40] A Kallioniemi O-P Kallioniemi D Sudar et al ldquoComparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumorsrdquo Science vol 258 no 5083 pp 818ndash821 1992
[41] D Pinkel and D G Albertson ldquoArray comparative genomichybridization and its applications in cancerrdquo Nature Geneticsvol 37 no 6 supplement pp S11ndashS17 2005
[42] Y Saillour M Cossee F Leturcq et al ldquoDetection of exoniccopy-number changes using a highly efficient oligonucleotide-based comparative genomic hybridization-array methodrdquoHuman Mutation vol 29 no 9 pp 1083ndash1090 2008
[43] J Staaf T Torngren E Rambech et al ldquoDetection and pre-cise mapping of germline rearrangements in BRCA1 BRCA2MSH2 and MLH1 using zoom-in array comparative genomichybridization (aCGH)rdquo Human Mutation vol 29 no 4 pp555ndash564 2008
[44] T A Peterson E Doughty andMG Kann ldquoTowards precisionmedicine advances in computational approaches for the anal-ysis of human variantsrdquo Journal of Molecular Biology vol 425no 21 pp 4047ndash4063 2013
[45] Y Choi G E Sims S Murphy J R Miller and A P ChanldquoPredicting the functional effect of amino acid substitutions andindelsrdquo PLoS ONE vol 7 no 10 Article ID e46688 2012
[46] S Stefl H Nishi M Petukh A R Panchenko and E AlexovldquoMolecular mechanisms of disease-causing missense muta-tionsrdquo Journal of Molecular Biology vol 425 no 21 pp 3919ndash3936 2013
[47] T G Kucukkal Y Yang S C Chapman W Cao and EAlexov ldquoComputational and experimental approaches to revealthe effects of single nucleotide polymorphisms with respect todisease diagnosticsrdquo International Journal ofMolecular Sciencesvol 15 no 6 pp 9670ndash9717 2014
[48] B M Kroncke C G Vanoye J Meiler A L George Jr andC R Sanders ldquoPersonalized biochemistry and biophysicsrdquoBiochemistry vol 54 no 16 pp 2551ndash2559 2015
[49] J Reumers J Schymkowitz and F Rousseau ldquoUsing structuralbioinformatics to investigate the impact of non synonymousSNPs and disease mutations scope and limitationsrdquo BMCBioinformatics vol 10 supplement 8 article S9 2009
[50] R Casadio M Vassura S Tiwari P Fariselli and P LuigiMartelli ldquoCorrelating disease-related mutations to their effecton protein stability a large-scale analysis of the human pro-teomerdquo Human Mutation vol 32 no 10 pp 1161ndash1170 2011
[51] D E V Pires J Chen T L Blundell and D B Ascher ldquoIn silicofunctional dissection of saturation mutagenesis interpretingthe relationship between phenotypes and changes in proteinstability interactions and activityrdquo Scientific Reports vol 6Article ID 19848 2016
[52] R Karchin M Agarwal A Sali F Couch and M S BeattieldquoClassifying variants of undetermined significance in BRCA2with protein likelihood ratiosrdquo Cancer Informatics vol 6 pp203ndash216 2008
[53] A A Yee A Savchenko A Ignachenko et al ldquoNMR and X-raycrystallography complementary tools in structural proteomicsof small proteinsrdquo Journal of the American Chemical Society vol127 no 47 pp 16512ndash16517 2005
[54] R D Finn P Coggill R Y Eberhardt et al ldquoThe Pfam proteinfamilies database towards a more sustainable futurerdquo NucleicAcids Research vol 44 no 1 pp D279ndashD285 2016
[55] S Hunter R Apweiler T K Attwood et al ldquoInterPro theintegrative protein signature databaserdquo Nucleic Acids Researchvol 37 no 1 pp D211ndashD215 2009
[56] J A Clapperton I A Manke D M Lowery et al ldquoStruc-ture and mechanism of BRCA1 BRCT domain recognition ofphosphorylated BACH1 with implications for cancerrdquo NatureStructural amp Molecular Biology vol 11 no 6 pp 512ndash518 2004
[57] S L Clark A M Rodriguez R R Snyder G D V Hankinsand D Boehning ldquoStructure-function of the tumor suppressorBRCA1rdquo Computational and Structural Biotechnology Journalvol 1 no 1 pp 1ndash8 2012
[58] R S Williams R Green and J N M Glover ldquoCrystal structureof the BRCT repeat region from the breast cancer-associatedprotein BRCA1rdquo Nature Structural Biology vol 8 no 10 pp838ndash842 2001
[59] R S Williams M S Lee D D Hau and J N M GloverldquoStructural basis of phosphopeptide recognition by the BRCTdomain of BRCA1rdquo Nature Structural amp Molecular Biology vol11 no 6 pp 519ndash525 2004
[60] P J E Rowling R Cook and L S Itzhaki ldquoToward classifica-tion of BRCA1 missense variants using a biophysical approachrdquoThe Journal of Biological Chemistry vol 285 no 26 pp 20080ndash20087 2010
[61] J N M Glover ldquoInsights into the molecular basis of humanhereditary breast cancer from studies of the BRCA1 BRCTdomainrdquo Familial Cancer vol 5 no 1 pp 89ndash93 2006