-
HIGH ABUNDANCE mtDNA RECOMBINANTS IN HETEROPLASMIC MICE:
Implications for AssistedReproduction and Therapeutic Cloning. D.C.
Wallace, M. Crimi. Center for Molecular and Mitochondrial
Medicineand Genetics (MAMMAG), University of California, Irvine,
Irvine, CA 92697-3940.
To determine if mammalian mtDNAs can recombine when mixed in the
same cytoplasm, we prepared mice usingfemale ES cell cybrids that
were heteroplasmic for the NZB and the Common inbred mtDNAs. These
mtDNAs differ in109 nucleotides (nt), including a BamHI
polymorphism at nt 4277. The two mtDNAs were maintained in
theheteroplasmic state through back crosses to C57BL/6 males. At
generations 12 through 15, the proportion of the twomtDNAs was
determined using the BamHI polymorphism, revealing three classes of
mice: those which maintainedmore than 30%, less than 30%, and no
detectable NZB mtDNAs. Heteroplasmy was confirmed in heart, brain,
liver,testis, muscle, ovarian, kidney, though it was highest in
liver. PCR amplification, cloning and sequencing of nt 3422 to5552
from liver and ovary mtDNAs revealed that 4/9 liver and 1/7 ovary
mtDNAs contained one of more crossoversbetween the NZB and Common
mtDNAs. This extraordinary finding was confirmed by purifying the
liver mtDNA froma heteroplasmic animal using two sequential
CsCl-ethidium bromide gradients, amplification, cloning and
sequencingand finding that 4 of 14 clones were recombinants, with
the sequence shifting from the Common mtDNA to the NZBmtDNA in two
distinct regions: nt 3600-4400 including ND1-3 tRNAs - ND2 and nt
8600-9400 encompassing COX III.A related analysis revealed
recombinant molecules in 1/10 heart, 1/14 brain, 5/12 kidney, 2/12
muscle, 0/6 testis, and4/14 ovarian clones. Somatic mutations were
also observed. Therefore, recombinant mtDNAs can be prevalent
inheteroplasmic mammals. These results are disturbing. In assisted
reproduction, the cytoplasm from younger oocytes isbeing injected
into older oocytes resulting in heteroplasmic children (Barritt,
J., et al., 2001, Hum Repro 16:513) and intherapeutic cloning
heteroplasmy can be expected. Consequently, both practices could
result in recombinant mtDNAs,something that evolution has gone to
considerable lengths to avoid. Supported from NIH grant NS21328
(DCW) andDept. Neurological Sciences, University of Milan,
fellowship (MC) .
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Development of a novel orally administered macrophage mediated
gene therapy for Gaucher Disease. E.I. Ginns,D.M. Faryna, M.
Galdzicka, C. Chrzanowski, G.R. Ostroff. Program in Medical
Genetics, Pediatric Neurology, UMassMedical School, Worcester, MA
01605.
A novel orally administered macrophage delivered gene therapy is
being developed for treatment of Gaucher disease.Current enzyme
replacement therapy (ERT) improves blood counts and reverses
hepatosplenomegaly. However, ERT isintravenously administered,
costly, and has not significantly addressed bone or neurological
complications. To evaluatethe efficacy of our orally administered
gene therapy method, DNA encoding human glucocerebrosidase
(huGBA)wasformulated inside micron-sized yeast cell wall particles
(YCWP). YCWP-huGBA formulations were used to introducehuGBA DNA
into J774 murine macrophages in culture and into long-lived Gaucher
mice generated by gene targeting.As observed in human patients, the
reduced GBA activity in these Gaucher mice results in
glucocerebroside storage andGaucher cells in tissues. The clinical
manifestations in these mice can be accelerated by short courses of
conduritol--epoxide. Following oral intake, YCWP-huGBA formulations
are taken up through intestinal Peyers Patches where theyare
phagocytosed by macrophages that then migrate to organs of the
reticuloendothelial system. Within the macrophageendosome the huGBA
DNA is released at acid pH, migrates to the nucleus, and is
expressed to produce normal huGBA.This macrophage mediated
treatment results in huGBA expression in J774 murine macrophages
in-vitro, and in tissuesof Gaucher mice. Preliminary findings
suggest that this therapy sufficiently corrects tissue GBA activity
to amelioratesymptoms in treated, compared to untreated, severely
affected Gaucher mice. As a consequence of improved delivery
ofhuGBA, we expect that this approach will achieve significant
reversal of tissue pathology, including bone. Ifmacrophages
containing huGBA migrate into brain, then resulting increased GBA
levels could also provide clinicalbenefit for neurologic
manifestations of the disease. The successful development of this
therapy should provide a safer,more efficient and cost effective
treatment for patients with Gaucher disease, as well as other
lysosomal diseases.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Adiponectin: An Inherited Survival Factor in Families with
Exceptional Longevity. A.R. Shuldiner1, J. Crandall2,T.I. Pollin1,
K. Tanner1, M. Rincon2, R. Lipton2, N. Barzilai2, G. Atzmon2. 1)
University of Maryland School ofMedicine,Baltimore; 2) Institute
for Aging Research AECOM, New York.
Adiponectin (ADIPOQ), a serum protein expressed by adipose
tissue, is lower with obesity and has a protective roleagainst
insulin resistance and atherosclerosis. To test the hypothesis that
ADIPOQ influences survival, we examinedADIPOQ levels and genetic
variation in the ADIPOQ gene in subjects with exceptional
longevity. We studied unrelatedAshkenazi Jews between the ages of
60 and 108 years (n=366), of which 169 where over the age of 95
years. To testheritability, ADIPOQ levels were also measured in 222
of their offspring. Homozygosity for the ADIPOQ +2019del/del
increased from 12% at age 65 to 31% at age 105, (=0.35, p=0.05) and
was associated with significantly higherserum ADIPOQ levels,
independent of BMI (p = 0.01). Serum ADIPOQ levels were greater in
subjects with exceptionallongevity (13.00.5 ug/mL vs. 17.1 0.6
ug/mL in age 60-94 years and 95 years, respectively, p=0.0001), an
effect thatwas independent of gender or BMI. Subjects with high
serum ADIPOQ levels had a greater percentage of large HDLand LDL
particle sizes(by NMR spectroscopy), higher HDL-cholesterol levels,
lower insulin resistance (by HOMA) andlower prevalence of the
metabolic syndrome (by NCEP III guidlines)compared to those with
low ADIPOQ levels. Thedistribution of adiponectin levels in the
offspring was bimodal, and ADIPOQ levels were significantly
heritable(h2=0.36, p=0.05) in offspring, suggesting inheritance of
a high (protective) ADIPOQ phenotype in families withexceptional
longevity. Exceptional longevity is associated with high levels of
ADIPOQ and with favorable lipoproteinprofile and protection from
the metabolic syndrome. This phenotype is inherited in part by a
novel ADIPOQ genotype.We suggest that ADIPOQ may be an inherited
survival factor in exceptional longevity acting by increasing
insulinsensitivity and providing protection from the metabolic
syndrome and cardiovascular disease.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Homozygous HOXA1 mutations disrupt human brainstem, inner ear,
cardiovascular, and cognitivedevelopment. M.A. Tischfield1,3, T.M.
Bosley4,5, M.A.M Salih6, I.A. Alorainy7, E.C. Sener8, M.J. Nester5,
D.T.Oystreck4, W. Chan1, C. Andrews1, R.P. Erickson9, E.C.
Engle1,2,3. 1) Dept of Medicine, Program in Genomics,Children's
Hosp Boston; 2) Dept of Neurology, Children's Hosp Boston; 3)
Program in Neuroscience, Div Med Sci,Harvard Med School, Boston; 4)
Neuro-ophthalmology Div, King Khaled Eye Specialist Hospital,
Riyadh, SA; 5)Neuroscience Dept, King Faisal Specialist Hospital
and Research Centre, Riyadh, SA; 6) Division of Peds Neuro,Depart
of Peds, College of Medicine, King Khalid Univ Hosp, Riyadh, SA; 7)
Depart of Radiology & DiagnosticImaging, College of Med, King
Saud University, Riyadh, SA; 8) Dept of Ophthal, Hacettepe Univ
Hospital, Ankara,Turkey; 9) Dept of Pediatrics, Mol. & Cell
Biology, Univ. of Arizona College of Med, Tucson, AZ.
Using a positional cloning approach, we have identified
homozygous coding mutations in HOXA1 in three geneticallyisolated
populations in the Middle East and American southwest. Affected
individuals have a pleiotropic spectrum ofphenotypes including
horizontal gaze abnormalities, facial weakness, deafness,
hypoventilation, skull deformities,autism, mental retardation,
internal carotid artery malformations, and conotruncal heart
defects. Although horizontalgaze abnormalities and sensorineural
deafness are the most penetrant aspects of the syndrome, the
remainingphenotypes demonstrate considerable variable expressivity
that can be dependent upon genetic background. This is thefirst
report of mutations in a HOX gene critical for CNS development and
the first description of viable homozygousmutations in any human
HOX gene. The identified mutations, two non-sense and one
frameshift, introduce prematurestop codons in exon 1. Truncated
mutant proteins should lack all known functional domains resulting
in a complete lossof HOXA1 function. Our results demonstrate a new
function for HOXA1 in vascular patterning that had not beenreported
in Hoxa1-/- mice. We also demonstrate that HOXA1 is necessary for
proper cognitive and behavioraldevelopment, intriguing considering
the canonical CNS expression domain of Hoxa1 has a rostal boundary
in thebrainstem.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Ancient Hybridization in the History of Homo sapiens. M.F.
Hammer, D. Garrigan, Z. Mobasher, S.B. Kingan, J.A.Wilder. Div
Biotechnology, Univ Arizona, Tucson, AZ.
Fossil evidence links human ancestry with populations that
evolved modern gracile morphology in Africa 130,000 -160,000 years
ago. Yet fossils alone do not provide clear answers to the question
of whether the ancestors of all modernHomo sapiens comprised a
single African population or an amalgamation of distinct archaic
populations. DNAsequence data have consistently supported a single
origin model in which anatomically modern Africans expanded
andcompletely replaced all other archaic hominin populations. Here
we present novel sequence data from a 17.5-kb X-linked locus,
Xp21.1, that exhibits ancient divergence among lineages. We analyze
levels of haplotype divergence andlinkage disequilibrium (LD) in
the framework of models predicting patterns of nucleotide variation
expected as aconsequence of admixture between historically isolated
subpopulations. No previous human polymorphism study hasbeen
specifically designed to utilize these measures to assess the
probability of a single population origin foranatomically modern
humans. The Xp21.1 locus was selected because it is located in a
non-coding region of the Xchromosome with moderately high levels of
recombination and very low gene density. This serves to minimize
thepotenital impact of natural selection acting on linked
functional sites. Monte Carlo computer simulations show that it
ishighly improbable that the pattern of nucleotide variation
observed at the Xp21.1 locus could arise in a single,
panmicticpopulation, as predicted by the recent African replacement
model. We consider several plausible alternative hypothesesand
conclude that ancient population structure in the evolutionary
lineage leading to AMH is the most likely explanationfor the Xp21.1
data. This inference supports human evolution models that
incorporate admixture between divergentAfrican branches of the
genus Homo.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
The International HapMap Project: A haplotype map of the human
genome with 4 million SNPs. P. Donnelly, onbehalf of The
International HapMap Consortium. University of Oxford, Oxford,
UK.
The International HapMap Consortium, which includes scientists
from Canada, China, Japan, Nigeria, the UK, and theUS, has
developed a human haplotype map describing the common properties
and patterns of DNA sequence variationacross the human genome. The
primary data are the genotypes of about four million SNPs, their
allele frequencies, andthe degree of association among them, in a
set of 270 DNA samples. In addition, ten 500 kb ENCODE regions
wereresequenced in 48 samples and the HapMap samples were genotyped
for all known and newly discovered SNPs. SNPswere genotyped using
technologies developed by Perlegen, Illumina, Third Wave, Sequenom,
ParAllele, and Perkin-Elmer. Four quality assessment exercises
found that the data are > 99% complete and > 99.5% accurate.
The HapMapdata are freely and publicly available at the HapMap Data
Coordination Center (www.hapmap.org and hapmap.jst.go.jp)and dbSNP
(www.ncbi.nlm.nih.gov/SNP), and are incorporated into the UCSC
Genome Browser and Ensembl. TheDNA samples were collected through
processes of community engagement and public consultation with
individualinformed consent. The 270 samples included 30 Yoruba
trios from Ibadan, Nigeria, 45 Japanese from Tokyo, Japan, 45Han
Chinese from Beijing, China, and 30 Utah CEPH trios (northern and
western European ancestry). Analyses of thedata were performed on
both the ENCODE genotypes and the genome-wide HapMap genotypes to
assess how sensitivethe inferences of patterns of LD and choice of
markers for tagging haplotypes are to SNP density. Specifically,
weanalyzed the types and patterns of LD and common haplotypes,
estimated a fine-structure recombination map, searchedfor
associations with genomic features, and sought evidence for natural
selection as an explanation of specific patterns.All analyses were
conducted genome-wide, for each population sampled.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Effect of mutation type and location on clinical outcome in 1081
Marfan syndrome patients or relatedphenotypes with FBN1 mutations :
an international study. L. Faivre1, G. Beroud2, A. Child3, B.
Callewaert4, C.Binquet1, E. Gautier1, E. Arbustini5, A.
Kiotsekoglou3, P. Comeglio3, C. Beroud2, C. Bonithon1, D.
Halliday6, C. Muti7,L. Ades8, J. De Baecker4, P. Coucke4, U.
Francke9, A. De Paepe4, G. Jondeau7, C. Boileau7. 1) Centre
dépidémiologie- investigation clinique et centre de Génétique, CHU
Dijon, France; 2) IURC, Montpellier, France; 3) St.
GeorgesHospital, London, UK; 4) Ghent University Hospital, Belgium;
5) Policlinico San Matteo, Pavia, Italy; 6) University ofOxford,
UK; 7) Hôpital Ambroise Paré, Boulogne, France; 8) Royal Alexandra
Hospital, Sydney, Australia; 9) StanfordUniversity Medical Center,
USA.
Mutations in the FBN1 gene cause Marfan syndrome (MFS) and have
been associated with a wide range of milderoverlapping phenotypes.
A large collaborative study based on the international FBN1
mutation database including 1081patients (820 probands) allowed us
to investigate the relationship between the FBN1 genotype and the
nature andseverity of the clinical phenotype. A set of qualitative
and quantitative clinical parameters was compared for
differentclasses of mutation. Patients with a FBN1 null mutation
had a more severe skeletal, aortic and lung phenotype thanpatients
with a mutation producing an altered protein (p=0.005, p=0.0004 and
p=0.007), whereas the frequency ofectopia lentis was lower
(p0.0001). A higher risk of ectopia lentis was found in patients
with a missense mutationsubstituting or producing a cysteine when
compared to patients with other missense mutations (p0.0001), as
well as witha mutation within the 5 region versus 3 (p=0.0003).
Patients with mutations within exons 24-32 had a more
severephenotype (younger age at diagnosis, at aortic dilatation or
aortic surgery and shorter survival) than patients withmutations
located elsewhere (p0.0001), even when neonatal MFS were excluded.
No significant differences were foundfor any clinical parameter in
patients with a mutation located in EGF domains versus TGFBP
domains. This study is thelargest ever reported and indicates
high-risk groups for cardiac and ophthalmologic manifestations, and
can be helpful inmonitoring patients with a FBN1 mutation.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
A meta-analysis of FXTAS patients with and without family
history of fragile X syndrome: a probable thresholdmodel for the
toxicity of CGG repeats. S. Jacquemont1, L. Beckett2, M. Leehey3,
F. Tassone4, R. Hagerman5, P.Hagerman3. 1) Département de
Génétique, C H U, Nantes, France; 2) Division of Biostatistics,
Department of PublicHealth Sciences, UC Davis, CA; 3) Department of
Neurology, University of Colorado Health Sciences Center Denver,CO;
4) Department of Biochemistry and Molecular Medicine, School of
Medicine, University of California Davis, CA;5) MIND Institute,
University of California Davis Medical Center Sacramento, CA.
Results of genetic screening for the FMR1 premutation (55-200
CGG repeats) have been published by 12 independentgroups. 2820
patients were ascertained through movement disorder clinics
independent of their family history. Weconducted a meta analysis in
order to determine the prevalence of FXTAS for the main movement
disorder diagnosesand established guidelines for fragile X
premutation screening in neurology clinics. The overall prevalence
of the FMR1premutation regardless of the movement disorder
diagnosis is 1%, 7 times greater than expected based on the
prevalenceof the premutation in males of the general population (OR
= 6.9 ; 95% CI : 1.8 , 65.5). This figure has limited
relevancesince the excess of premutation allele was significant
only in the group of males presenting cerebellar ataxia after
theage of 50 years. In that group, the prevalence of the
premutation rises to 2.7% ; (21/778), 22 times greater than
expectedbased on the prevalence of the premutation in males of the
general population (OR = 22.4 ; 95% CI : 5.8, 86.7). Theclinical
descriptions of these patients were compared to the cases
identified through fragile X families. The metaanalysis of the
molecular data from patients recruited with or without a family
history of fragile X supports thehypothesis of a threshold model of
the toxicity of CGG repeats. We conclude that the toxicity of CGG
repeats below 70may not have functional repercussions or may be
associated with FXTAS related disorders. This meta-analysis
hasconsequences for the projected prevalence of FXTAS in the
general population, which is lower than the initial estimateof 1/
3000.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
FXTAS: a descriptive study of premutation carriers from fragile
X families. E.G. Allen, J. Juncos, M. Rusin, G.Novak, L. Shubeck,
S.L. Sherman, R. Letz. Emory Univ, Atlanta, GA.
We are conducting a study to further examine the symptoms,
penetrance, and risk factors associated with thetremor/ataxia
syndrome (FXTAS) among carriers of premutation alleles of the FMR1
gene. Our study populationincludes all siblings of premutation
carrier males over the age of 50 identified through a survey of
families with fragileX syndrome. We conducted a comprehensive
battery of tests including a medical history, a neuropsychological
testbattery, and quantitative neurological assessment. Within the
neurological assessment, we use a series of tests to
obtainobjective, quantitative measures of key features observed in
FXTAS cases to date: 1) resting, postural, and intentionupper-limb
tremor, 2) reduced vibration sensation, a surrogate of possible
neuropathy, 3) decreased postural stability,and 4) dysdiadokinesis,
or upper limb uncoordination. We have obtained these measures on 49
males (mean age=63.8)and 20 females (mean age=65.3). Subjects were
scored as expressing the above phenotypes if they scored above
the95th percentile of age-adjusted standards. Among premutation
males, we have seen a significantly increased incidenceof tremor
(45%), peripheral neuropathy (14%), decreased postural stability
(57%), and dysdiadokinesis (78%). Incontrast, these phenotypes were
not markedly increased in females or non-carrier male relatives.
These findings as wellas neuropsychological measures will be
examined with molecular correlates of FMR1.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Spectrum of CHD7 mutations in 113 patients with CHARGE Syndrome.
S.R. Lalani1, A.M. Safiullah1, S.D.Fernbach1, L.M. Molinari1, C.
Bacino1, S.L. Davenport4, M.A. Hefner2, J.M. Graham Jr.3, J.W.
Belmont1. 1)Department of Molecular and Human Genetics, Baylor
College of Medicine, Houston, TX; 2) Department of Pediatrics,St.
Louis University School of Medicine St. Louis, MO; 3) Medical
Genetics Institute, Cedars-Sinai Medical Center,Los Angeles, CA; 4)
Sensory Genetics/Neuro-Development, St. Paul, Minnesota.
CHARGE is a complex birth defect, characterized by non-random
occurrence of coloboma, choanal atresia, cranialnerve defects,
distinctive inner ear abnormalities, heart and urogenital anomalies
and growth retardation. Recently,intragenic mutation of CHD7, the
chromodomain-helicase DNA binding protein was reported to be a
major cause ofCHARGE syndrome. Chromatin remodeling is a
well-recognized mechanism of gene expression regulation and thegene
is likely to play a significant role in embryonic development and
cell cycle regulation. We report the spectrum ofCHD7 mutations in
113 individuals with sporadic and familial CHARGE. Mutations were
found in 62 patients (55%)distributed throughout the gene. About
68% were truncating mutations, most likely leading to
haploinsufficiency. Wehave obtained phenotypic information on all
patients and have performed multivariate analysis conditioned on
themutation status. The analysis shows that congenital heart
defects (50/55 compared to mutation negative group, p=0.014)and
growth retardation (42/48 compared to mutation negative group,
p=0.002) are more frequent in patients with CHD7mutation. Mouse
embryo in situ hybridization shows expression of this gene in the
oto-acoustic complex, brain, ear,pharyngeal endoderm, and heart
tube. We have also performed microarray analysis of gene expression
usinglymphoblastoid cell lines of seven CHD7 mutation positive
patients and compared the expression pattern in these to
fivecontrol subjects and four affected patients with no abnormality
of CHD7 gene by sequence analysis. Preliminaryanalysis shows
significant gene expression differences among these three groups.
The differential gene expressionpattern in the CHD7 mutation
negative group may further assist in understanding the molecular
basis of disorder in thisgroup of affected children.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Phenotypic spectrum of CHARGE syndrome with CHD7 mutations. K.
Kosaki1, M. Aramaki1, T. Udaka1, R.Koaki2, Y. Makita3, N. Okamoto4,
H. Yoshihashi5, H. Oki5, Y. Fukushima6, H. Kawame7. 1) Dept
Pediatrics, Div MedGen, Keio Univ, Tokyo, Japan; 2) Dept of Clin
Genet & Mol Med, Nat'l Children'fs Med Ctr, Tokyo, Japan; 3)
Dept ofPediatrics, Asahikawa Med Col, Asahikawa, Japan; 4)
Department of Planning and Research, Osaka Medical Centerand Res
Inst for Maternal and Child Health, Osaka, Japan; 5) Tokyo
Metropolitan Kiyose Children'fs Hosp,Tokyo; 6)Dept of Med Genet,
Shinshu Univ Sch of Med, Matsumoto, Japan; 7) Division of Clinical
Genetics, Nagano Children'sHosp, Nagano, Japan.
CHARGE syndrome is characterized by a constellation of
non-randomly associated malformations: C - coloboma ofthe iris or
retina, H - heart defects, A - atresia of the choanae, R -
retardation of growth and/or development, G - genitalanomalies, and
E - ear abnormalities. Recently, Vissers et al. identified the gene
Chromodomain helicase DNA-bindingprotein-7 (CHD7) at chromosome
8q12.1 as a causative gene of CHARGE syndrome. We further
delineated thephenotypic spectrum of CHARGE patients with mutations
in CHD7. Twenty-three patients who fulfilled Blake'fscriteria were
screened for CHD7 mutations by using DHPLC. PCR products
corresponding to all variant elution profileswere sequenced
bidirectionally using the dideoxy sequencing method. We identified
heterozygous CHD7 mutations in17 (71%) of the 24 patients enrolled
in the study: 7 frameshift mutations, 6 nonsense mutations, 3
splice-site mutations,and 1 intragenic deletion from exon 8 to 12.
These mutation classes are likely to result in a prematurely
truncatedprotein or the loss of protein expression. Altogether, 15
of the 17 cases had coloboma of the eyes. However, only 2patients
had iridal coloboma and 5 patients had only disc coloboma, without
iridal or retinal coloboma. Thus, the retinaand optic disc must be
thoroughly examined when CHARGE syndrome is suspected, even in the
absence of iridalcoloboma. Choanal atresia/stenosis was not common
(5 patients) and was less frequent than oral clefts (8
patients).Severe hearing-loss, laryngomalacia, and developmental
delays were prevalent. Genetic testing of CHD7 will be helpfulin
confirming the diagnosis and in providing accurate genetic
counseling to the families.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Genotype-phenotype correlations in Rubinstein-Taybi syndrome.
E.K. Schorry1, M. Keddache1, B. Abiramikumar1,N. Lanphear2, J.
Rubinstein2, G.A. Grabowski1. 1) Div. of Human Genetics; 2) Div. of
Developmental Disabilities,Cincinnati Children's Hospital.
Rubinstein-Taybi syndrome (RTS) is a well-described multiple
congenital anomaly/mental retardation syndromewhich usually occurs
sporadically. Loss of function in CREBBP or EP300 genes, which
encode histone acetyltransferases, has been found in about 50% of
patients with RTS, with the majority of mutations in CREBBP.
Weperformed mutation analysis of CREBBP in 98 patients who met
diagnostic criteria for RTS during 2 international RTSfamily
conferences. DNA was extracted from peripheral blood, and mutation
analysis of CREBBP was performed on all31 coding exons and
exon-intron junctions by bidirectional comparative PCR sequencing.
A total of 59 differentvariations were observed in the DNA
sequence. All mutations involving a change in the protein were
unique, and wereequally distributed throughout the 31 coding exons.
Twenty-five of the mutations created a truncated protein product
orclearly altered the splice-donor/splice-acceptor consensus
sequence; 11 mutations resulted in single amino acid
changes.Forty-five patients did not have an identifiable mutation
in CREBBP. Extensive phenotypic data was also collected on the
patients during the conference. All patients had the
characteristicfacies, broad thumbs, and mental retardation. We
analyzed phenotypic data to determine correlations with
specificmutation types, i.e., truncating, splice site, single amino
acid substitutions, or no identifiable mutation. There were
nodifferences in the facial and broad thumb phenotype between the 4
groups, with all groups displaying the characteristicdysmorphology.
Degree of mental retardation and growth retardation were similar in
all groups. Congenital heartdisease was seen in 40% of the group
overall, but was less frequent (18%) in patients with single amino
acidsubstitutions. Interestingly, duplication of the great toe was
seen in 8 patients with no identifiable mutation, but in noneof the
patients with CREBBP mutations. Further research is needed to
determine the additional genes which cause thisphenotype.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Phenotypic characterization of familial oculo-auriculo-vertebral
(OAV) spectrum: assessment of 19 additionalfamilies. A.E. Beck1, L.
Hudgins2, A.W. Grix4, N.H. Robin5, E. Chen6, L.C. Lazzeroni3, H.E.
Hoyme2, U. Francke1,2. 1)Dept Genetics and; 2) Dept Pediatrics, Div
Medical Genetics and; 3) Dept Health Research & Policy, Div
Biostatistics,Stanford Univ, Stanford, CA; 4) Permanente Medical
Group, Sacramento, CA; 5) Dept Genetics, Univ Alabama atBirmingham,
Birmingham, AL; 6) Permanente Medical Group, Oakland, CA.
Although the OAV spectrum is considered a sporadic condition, a
few large studies suggest that up to 45% ofprobands with hemifacial
microsomia or microtia have affected relatives. In the present
study, we evaluated 19 familiesin which probands with OAV had
additional family members with OAV features. We ascertained 54
affected familymembers (including probands) and 18 normal relatives
who were apparently transmitting carriers. Of thosephenotypically
affected, 57% were female and 43% were male. 63% had unilateral and
37% had bilateral findings. 93%exhibited auricular anomalies
including preauricular tags (23 cases) and microtia (30 cases,
often with canalnarrowing/atresia and conductive deafness). Eye
malformations in 20% included epibulbar dermoids
(8/11),microphthalmia, and iris/retinal colobomas. Mild facial
asymmetry was seen in 30%, but the marked hemifacialmicrosomia
often seen in sporadic OAV was uncommon. Renal (4%), vertebral
(11%) and cardiac (7 %) anomaliesoccurred less frequently than in
sporadic cases. Only 2 affected subjects had intellectual
impairment. Chromosomalstudies in probands were normal. Although
mouse models and candidate regions in the human genome are
underinvestigation, no human gene mutations for OAV are known. Our
new data suggest that there are genetic subtypeswithin the OAV
spectrum consistent with autosomal dominant and possibly X-linked
inheritance of a single genedisplaying variable expressivity and
incomplete penetrance. Careful evaluation of relatives for minor
manifestations ofOAV is essential before quoting recurrence risks.
Since microtia can result from aberrant migration of neural crest
cellsinto the 1st/2nd branchial arches during embryonic weeks 4-5,
it is tempting to postulate that the altered gene(s) in ourfamilial
cases are involved with neural crest regulation.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Array CGH identifies chromosome abnormalities with unexpected
clinical variability in contiguous genesyndromes. B.A. Bejjani,
B.C. Ballif, C.D. Kashork, E. Rorem, K. Sundin, L.G. Shaffer.
Signature GenomicLaboratories, LLC, Spokane, WA.
Array CGH provides distinct advantages over conventional
cytogenetics by detecting aneuploidy,
microdeletions,microduplications, and subtelomeric rearrangements
in a single, simultaneous assay. The SignatureChip was designed
todetect the common microdeletions, reciprocal microduplications,
subtelomeric and pericentromeric alterations,unbalanced
translocations, and aneuploidy, while avoiding most common
population variants. The design also allowsfor distinguishing the
common-sized microdeletions that are flanked by low copy repeats
from larger deletions throughthe use of flanking control loci. The
array uses 831 BACs covering 126 clinical and 104 control loci in
3-6 clonecontigs. Of the 1,300 clinical cases submitted to our
laboratory for diagnostic testing, we identified 73 cases (5.6%)
withclinically relevant abnormalities. The majority of these cases
have had at least one previous cytogenetic study. Amongthese, 22
telomeric deletions or unbalanced rearrangements and 21 cases with
syndromic deletions were identified. Wealso identified 3 cases of
interstitial deletions of 1p36 and three cases of interstitial
duplications on 3q, 4q and 16q, noneof which would be detected
using available sets of subtelomere probes. The finding of 21 cases
with syndromicmicrodeletions most of which were submitted by
clinical geneticists after at least one previous cytogenetic
studysuggests that these well-characterized syndromes have
unrecognized variability in clinical presentation that may
preventthe clinician from readily reaching a diagnosis. The
clinical variability of these syndromes will be reviewed. In
addition,microduplications of 15q12, 17p11.2, and 22q11.2 have been
identified, confirming that reciprocal products of
commonmicrodeletion syndromes will be detected. These
microduplications are unlikely to be identified by
conventionalmetaphase FISH assays. Thus, array CGH is a powerful
approach for uncovering subtelomeric rearrangements,microdeletions
and microduplications, even in patients who are not suggestive of a
particular syndrome.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
9q34.3 microdeletion syndrome: clinical and genetic insights.
S.A. Yatsenko1, H. Firth4, S. Tomkins5, O. Rittinger6,E. Lammer7,
K.S. Lewis8, S.W. Cheung1, P. Stankiewicz1, J.R. Lupski1,2,3. 1)
Dept Molec & Human Genetics, BaylorCol Medicine, Houston, TX;
2) Pediatrics, Baylor College of Medicine; 3) Texas Children
Hospital, Houston, TX; 4)Addenbrookes Hospital, Cambridge, UK; 5)
St. Michael's Hospital, Bristol, UK; 6) Klinische
Genetik,Landeskinderklinik Salzburg, Salzburg, Austria; 7)
Children's Hospital and Research Center at Oakland, Oakland, CA;8)
St. Joseph's Hospital And Medical Center, Phoenix, AZ.
9q34.3 microdeletion syndrome is a contiguous gene syndrome
characterized by craniofacial dysmorphism, neonatalhypotonia,
childhood obesity, microcephaly, mental retardation and absence of
expressive speech. Depending on thedeletion size, additional
clinical features can include congenital heart defects, seizures,
abnormal male genitalia, limband brain anomalies. Recently we
identified an ~700 kb critical region at the most distal portion of
9q34.3encompassing genes EHMT1 (euchromatic histone
methyltransferase) and CACNA1B in association with a
minimalphenotype (Yatsenko et al. 2005). Haploinsufficiency of
EHMT1 in all our patients and disruption of EHMT1 identifiedin a
child with features of 9q deletion syndrome (Kleefstra et al. 2005)
provide the first evidence that dysregulation ofhistone
modifications and chromatin structure can be a pathophysiological
mechanism underlying mental retardationand neurobehavioral
disorders. Here we present the FISH mapping of the breakpoints in
six new patients with del(9)(q34.3) using BAC and fosmidclones, and
compare the deletion sizes and clinical manifestations. We
constructed a deletion map of 9q34.3chromosome region using cell
lines from eleven affected individuals. In addition molecular
cytogenetic analysisrevealed one patient with
der(9)del(9)(q34.3)dup(9)(q34.3q34.2). Our investigation showed
that the 9q34.3 deletionsvaried from ~700 kb to 3.5 Mb in size and
complexity of the rearrangement can be detected only by FISH or
microarrayCGH analyses. These observations provide new insights for
genetic counseling and for the search of new candidategenes. We
discuss genotype-phenotype correlation, and the possible molecular
mechanism of deletion.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Overgrowth in association with chromosomal anomalies. G.S. Ball,
R.K. Aldrich, J. Lee, S. Li, K.A. Casas.Department of Pediatrics,
University of Oklahoma, Rare Chromosomal Anomalies Registry,
Oklahoma City, OK.
Overgrowth is a feature of very few chromosomal syndromes. We
report two new patients with overgrowth. Patient 1presented at
birth with distal joint contractures and failed newborn hearing
screen. Imaging studies revealed agenesis ofthe corpus callosum and
left cystic kidney. Routine chromosome analysis performed shortly
after birth showed extramaterial on chromosome 6q. Subtelomeric
fluorescent in situ hybridization (FISH) using a 10p subtelomeric
probeconfirmed that the extra material on 6q was derived from 10p,
resulting in terminal deletion of 6q and trisomy 10p. At 5months of
age, her length was 4 standard deviations above the mean for age.
Weight and occipitofrontal circumferencewere at the 75th centile.
Palm length was greater than the 97th centile for age, and there
was an advanced bone age of 2years. Patient 2 presented at 15 years
of age with mild mental retardation and recent onset of seizures
and hypertension.Her height was over 3 standard deviations above
the mean for an adult female, weight was 4 standard deviations
abovethe mean for an adult female, and occipitofrontal
circumference was 3 standard deviations above the mean for an
adultfemale. Palm length was at the 97th centile for an adult
female. Routine chromosome analysis had been performed inearly
childhood with normal results. Subtelomeric FISH testing at 15
years of age revealed an unbalancedrearrangement with deletion of
8p and trisomy of 12p. Overgrowth is not a reported feature of the
6q deletion or 10pduplication syndromes, suggesting a fusion gene
or other mechanism unique to the (6;10) translocation in Patient
1.Overgrowth is not a reported feature of the 8p deletion syndrome,
but increased birth length and birth weight havepreviously been
associated with duplication 12p and tetrasomy 12p, suggesting a
dosage effect of a growth factor gene.Results in Patient 2 narrow
the candidate region for this gene(s) to the 12p subtelomere.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Identification of Segmental Chromosomal Duplications and
Deletions in Prostate Cancer. A. Pearlman, M.Salman, H. Ostrer.
Human Genetics Program, New York University School of Medicine, New
York, NY.
Recurrent somatic genetic alterations contribute to the
development and progression of prostate cancer. These eventscan be
identified using arrayCGH, a high-resolution technique that maps
the duplicated or deleted segments onto arraysof well-characterized
bacterial artificial chromosomes (BACs) or oligonucleotides. As
part of our efforts to identifysomatic alterations that are
predictive of racial differences or prognosis, and to aid the
identification of tumor suppressorgenes, we compared the
performance of Affymetrix 10K SNP chips and 19K BAC arrays using
three Gleason 7 prostatecancers and their matched normal tissue. We
analyzed the Affymetrix 10K SNP chip arrays using GDAS 2.0,dChipSNP
and the Affymetrix copy number tool and the BAC arrays using the
CGH Explorer, DNAcopy, vMAP, andGLAD segmentation tools. SNP chips
provided calls for 93-98% of the probes. The gender calls were as
expected for allsamples and the concordance between replicates was
97%. No duplication events were identified with confidence.
LOHevents were observed on chromosomes 1, 13, and 21 for tumor 3
and in an overlapping region on 13 for tumor 1.Among the 60,000
events sampled, 4 small, LOH events occurred in single samples, but
not in their replicates,suggesting that the frequency of false
positives was
-
Finding prostate cancer susceptibility genes using linkage-based
and candidate gene approaches. E.A. Ostrander,on behalf of The
International Consortium of Prostate Cancer Genetics. Cancer
Genetics Branch, NHGRI/NIH,Bethesda, MD.
Identification of prostate cancer (PC) susceptibility genes has
been hampered by the phenotypic heterogeneityassociated with the
disease, the large number of contributing loci, and the associated
difficulties in accurately stratifyingfamilies based on features of
disease, family history, or clinical presentation. Although
genome-wide screens have beenperformed in over a dozen independent
studies, few chromosomal regions have been consistently identified
as regionsof interest. The International Consortium for Prostate
Cancer Genetics (ICPCG) was formed to facilitate theidentification
of PC susceptibility genes. The ICPCG has undertaken analyses aimed
at utilizing the large familyresources available (over 1800 PC
families). These include: 1) combining genome-wide screen linkage
data from a totalof 1,233 PC families collected by members of the
ICPCG. Using parametric and non parametric analyses five
regionswere identified with suggestive linkage (LOD score >1.86)
on 5q12, 8p21, 15q11, 17q21 and 22q12, as well as asignificant
linkage at 22q12 (LOD 3.57) in a subset of families with 5 or more
affected members. These findings areconsistent with the hypothesis
of multiple PC susceptibility genes with modest effects, or several
major genessegregating in small subsets of families; 2) multi-locus
analysis of genome wide scan data in all families using
orderedsubset analysis and pedigree covariate analysis using
recursive partitioning; 3) a genome wide scan for linkage in
189families that have three or more first degree relatives with
aggressive prostate cancer; 4) genome wide scans for linkagein
families with both prostate and other cancers (e.g. gastric cancer)
and prostate cancer families of Ashkenazi Jewishdescent. Finally,
we have undertaken replication studies for previously described
loci on the X chromosome, and testingof provocative candidate genes
such as CHEK2 and NBS1. In addition, novel methodologic approaches
are beingdeveloped and implemented to analyze this large and
informative data set.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Genome-wide linkage scan for prostate cancer susceptibility loci
in men with aggressive disease: The Universityof Michigan Prostate
Cancer Genetics Project. L.A. Ho1, J.B. Dimmer2, Y. Wang1, E.M.
Gillanders3, K.A. Cooney2,E.M. Lange1. 1) University of North
Carolina, Chapel Hill, NC; 2) The University of Michigan, Ann
Arbor, MI; 3)NHGRI, NIH, Bethesda, MD.
Identifying prostate cancer susceptibility genes using linkage
analysis has proven to be a difficult challenge. It is clearthat
there is considerable genetic heterogeneity which results in
reduced power to detect linkage signals. Another likelycomplication
in mapping prostate cancer susceptibility genes is that there is
considerable heterogeneity in the phenotypeof prostate cancer, with
a considerable range in the aggressiveness of the disease even
among family members. Weperformed a genome-wide nonparametric
linkage scan for genes that predispose to aggressive forms of the
diseasebased on 405 genetic markers and 79 informative pedigrees
with two or more cases of aggressive disease. Onlyindividuals
defined to have aggressive disease were coded as "affected" in our
analyses. An indicator variable foraggressive disease was created
using the following criteria: 1) Regional or distant stage (based
on pathology if radicalprostatectomy done, otherwise clinical
stage, T3, T4, N1, M1) or 2) Gleason score of 7 or higher or poorly
differentiatedgrade (if no Gleason score is available) or 3) PSA at
diagnosis 20 or higher or 4) prostate cancer listed as primary
causeof death on death certificate. We found strong evidence for
linkage on chromosome 15q (LOD = 3.43) near markerD15S1002.
Suggestive linkage was found on chromosome 6p (LOD = 2.21) near
marker D6S422. Additional LODscores greater than 1 were found on
chromosomes 3q (1.43), 18q (1.17), 20q (1.05) and Xp (1.02). Using
a more rigiddefinition of prostate cancer will result in a severe
reduction in the effective sample sizes available for linkage
analysis.However, the reduction in heterogeneity or "noise" in the
phenotype may ultimately prove to increase the statisticalpower to
detect susceptibility genes for this complex trait.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Single locus gene mutations and prostate cancer: Can mutations
in the androgen receptor gene be directly linkedto the occurrence
of prostate cancer? B. Gottlieb1,3,5, C. Alvarado1, L.K. Beitel1,2,
K. Sircar4, A. Aprikian4, M.Trifiro1,2,3. 1) Dept Cell Genetics,
Lady Davis Inst Medical Res, Montreal, PQ, Canada; 2) Dept
Medicine, McGillUniversity, Montreal, PQ, Canada; 3) Dept Human
Genetics, McGill University, Montreal, PQ, Canada; 4) DeptUrology,
McGill University, Montreal, PQ, Canada; 5) Dept Biology. John
Abbott College, Ste Anne De Bellvue, PQ,Canada.
One of the Holy Grails of cancer genetics has been to find
mutations in a specific gene linked directly to a specificcancer. A
possible candidate has been the androgen receptor gene (AR), which
recent evidence has shown is importantin all prostate cancer (CaP)
stages. However, numerous studies that have examined the possible
association between ARmutations and CaP have produced very
inconclusive results. To resolve this issue we have considered the
fact that tumortissue heterogeneity has rarely been considered an
indicator of genetic heterogeneity. To see if AR genetic
heterogeneitydoes indeed exist we have conducted a micro-genetic
examination of tumors from CaP patients using laser
capturemicrodissection (LCM) to isolate specific
pathologically-identified tissues and then sequence their AR. To
help correlatemutations with tissue type, we have prepared
cancerous and benign tissue micro-arrays from prostates from men
withCaP. To date we have found a number of different mutations, in
distinct areas isolated from a single prostate, all ofwhich appear
to be unique. Mutations were found in both cancerous and benign
tissue from the central zone of theprostate, but not from benign
peripheral and transitional zones. Even more intriguing is that
different mutations werefound in cancerous and benign tissues.
Thus, it appears that somatic mosaicism of the AR occurs within CaP
tumors,although the exact relationship between AR mutations and CaP
must await an analysis of many more prostate tumors.We believe that
this study, which is the first to conduct a micro-genetic analysis
of AR mutations in CaP by utilizingLCM, points the way to an
approach that can much more accurately monitor genetic events that
lead to the initiation andprogression of cancer.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Allele-specific expression in familial pancreatic cancer. J.
Fan1, A. Maitra2, M. Bibikova1, E. Chudin1, E.W.Garcia1, D.
Barker1, P. Chen2, C. Karikari2, M.G. Goggins2, R.H. Hruban2, A.
Chakravarti2. 1) Illumina, Inc, SanDiego, CA; 2) Johns Hopkins
University, Baltimore, MD.
Pancreatic cancer afflicts over 30,000 individuals each year in
the United States, and is mostly lethal within months ofdiagnosis.
About 10% of pancreatic cancer is familial due to an inherited
predisposition. Despite considerable effort,germline mutations have
been identified in less than 20% of families, in a few genes such
as BRCA1/2. To enhancegene discovery, we analyzed germline
allele-specific expression (ASE) patterns in familial pancreatic
cancer (FPC). Ourcentral hypothesis is that the gene responsible
for FPC would be hemizygously mutated in the germline of
affectedindividuals, and hence, would only or predominantly be
expressed hemi-allelically from the wild-type allele (at least fora
subset of mutations such as nonsense mutations or deletions). A
microarray-based ASE assay was developed for 2,140coding SNPs
(cSNPs) derived from 680 cancer-associated genes, such as cell
cycle checkpoint and apoptosis-inducinggenes. Lymphoblastoid cell
lines from 48 age-matched individuals (16 CEPH family trios) were
analyzed for genomicheterozygosity and allele-specific differences
in expression, if any, at heterozygous cSNPs. Genes
demonstratingsignificant ASE in these normal individuals were
catalogued for reference. Subsequently, lymphoblastoid cell lines
from109 individuals with FPC were examined for ASE variations.
Overall, 331 genes demonstrated potential ASE in at leastone of the
157 samples, of which 201 demonstrated ASE in both CEPH and FPC
samples. These include knownimprinted genes as well as multiple
X-chromosome transcripts, confirming the overall validity of our
platform. Theremaining 130 genes demonstrated ASE in one or more of
the 109 FPC patient samples only, and include both knowngenes
implicated in FPC (e.g., BRCA1/2) as well as novel candidate genes
(e.g., XRCC5, BPAG1, AGO, etc.). We arecurrently validating
array-based ASE on a gene-by-gene basis using quantitative RT-PCR.
Mutational screening will beperformed for genes that display
confirmed ASE variation in FPC patients but not (or at a
significantly lower frequency)in CEPH controls.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
PRKAR1A mutations leading to Carney complex, PPNAD and related
disorders: Analysis of the largest databaseto date, identification
of large gene rearrangments and other novel disease-causing
alterations, and functionalcharacterization of expressed mutants.
A. Horvath, S. Boikos, F. Weinberg, E. Meoli, S. Stergiopoulos, T.
Bei, L.Matyakhina, I. Bossis, C. Stratakis. Section on
Endocrinology & Genetics, DEB, NICHD, NIH, Bethesda, MD.
Carney complex (CNC) and primary pigmented nodular
adrenocortical disease (PPNAD) are caused by inactivatingmutations
in PRKAR1A, the main regulatory subunit of cAMP-dependent protein
kinase (PKA). We have screened forPRKAR1A sequence changes a total
of 655 individuals from 171 non-related affacted kindreds, by DHPLC
and genomicsequencing; negative samples were then screened by
Southern blotting; 161 patients have been found to have 55
(33novel) different disease-causing single base substitutions or
small insertions/deletions scattered over the codingsequence and
the splice junctions. Six index cases have shown alternative
patterns in heterozygote state by Southernanalysis using EcoRI;
further analysis confirmed two distinct deletions, one ~5 kb from
the exon3/intron3 region, andanother deleting ~4 kb from between
exons 4 and 8. These data were confirmed by long distance PCR; FISH
did notshow any abnormalities. Four specimens were sharing common
EcoRI altered restriction pattern, suggesting mutationsof EcoRI
site. In contrast to previous studies, where the vast majority of
the mutations were proven to lead to NMD, andsubsequently, not to
be expressed, we identified 10 different expressed mutations, which
allowed us to assess thegenotype-phenotype correlations. Two of the
expressed mutations were associated with particularly aggressive
clinicalphenotype, two other were associated almost exclusively
with PPNAD only. We conclude that (1) there is a largenumber of new
PRKAR1A mutations associated with CNC and PPNAD, (2) the existence
of large gene rearrangementsconfirms the complex PRKAR1A
disease-mutation spectrum (3) approximately 20% of the mutations do
not undergoNMD and lead to the expression of various mutant PRKAR1A
proteins, and (4) there is genotype-phenotypecorrelation. This
study has important implications for counseling, and for the
molecular pathophysiology of PKA-associated tumorigenesis.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
A mouse model employing the tet-off system in an antisense
construct to Prkar1a confirms tTA-independentexpression of the
transgene and develops tumors consistent with Carney complex
(MIM160980). K.J. Griffin, F.Weinberg, S. Stergiopoulos, L.
Matyakhina, L. Kirschner, C. Stratakis. SEGEN, DEB, NICHD,NIH,
Bethesda, MD.
Carney complex (CNC) is a multiple neoplasia syndrome associated
with adrenal, pituitary, thyroid, and gonadalneoplasms, caused by
PRKAR1A mutations inactivating the type IA regulatory subunit of
protein kinase A (PKA). Atransgenic line carrying an antisense
transgene for exon 2 (X2as) of Prkar1a was crossed with a
tTA-transgene-expressing line. The resulting tTA/X2as mouse
mimicked CNC in many respects (Cancer Res 2004;64:8811, J MedGenet
2004;41:923), and developed lymphoid hyperplasia. The mice with
only the X2as transgene also developedabnormalities that were not
present in mice carrying tTA alone or in wild type mice of the same
background. Previousreports have suggested that the tet-off system
is leaky, allowing for the occasional expression of the target
constructwithout the tTA factor. This study reports 136 mice: wild
type (16), those carrying only the X2as transgene (60),
thoseexpressing tTA only (15), and the tTA/X2as (45) line published
recently. The X2as mice developed a variety of tumorsthat were not
present in controls, and not different from those in tTA/X2as mice.
These include follicular adenomas ofthe thyroid, pigmentation and
x-zone vacuolization of the adrenal cortex, uterine cysts, and
abnormalities of the testis.The X2as mice did not develop lymphoid
hyperplasia or lymphoma as frequently as tTA/X2as mice. PKA
activities inX2as tissues were intermediate between tTA/X2as and
controls. Genetic analysis of X2as-derived sarcomas
showedinvolvement of the mouse chromosome 11 Prkar1a locus.
Expression of the X2as transgene was confirmed by proteinand mRNA
studies; proliferation assays of X2as and tTA/X2as mouse-embryonic
fibroblasts were similar. We concludethat the X2as mice developed
tumors similar to those seen in CNC, tTA/X2as, and Prkar1a+/-mice.
These studiesconfirm the leakiness of the tet-off system but also
provide the opportunity to test which features are related to
Prkar1a-down regulation vs. expression of the tTA transgene.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Mutations and Interstitial Deletions Involving Exon 3 of the
Beta-catenin Gene are Detected in SporadicAdrenocortical Tumors. I.
Bourdeau, A. Lampron, M. Tadjine. Division of Endocrinology,
Department of Medicine,Centre hospitalier de lUniversité de
Montréal (CHUM)-Hôtel-Dieu, Montréal, Québec, Canada H2W 1T8.
Adrenocortical lesions are diagnosed frequently as benign
adenomas or nodular hyperplasias, and more rarely asmalignant
adrenocortical carcinomas. The genetic background of sporadic
adrenocortical lesions is poorly characterized.In our previous
work, involving large-scale cDNA microarray analysis, we
demonstrated aberrant differentialexpression of a number of
Wnt/-catenin signaling-related genes implicated in adrenocortical
hyperplasias. To furtherexplore the role of Wnt/-catenin signaling
in adrenocortical tumorigenesis, we conducted a search for
mutations of exon3 of the -catenin gene. DNA was extracted from 48
human adrenocortical samples and the human adrenocortical
cancercell line NCI-H295R. The adrenal tissue samples included 28
patients with adenomas, 4 with carcinomas, 13 withACTH-independent
macronodular adrenal hyperplasia (AIMAH) and 3 with ACTH-dependent
adrenal hyperplasia. Thesamples were screened for somatic mutations
in exon 3 of the -catenin gene using deletion screening by
polymerasechain reaction (PCR) and direct sequencing. We found
genetic alterations in 6 out of 28 adenomas (21,4%). There wereno
mutations detected in adrenocortical carcinomas, AIMAH,
ACTH-dependent hyperplasia and the NCI-H295R cellline. Three point
mutations occurred at potential serine/threonine phosphorylation
residues of codons 37 or 45 [S37C(n=2) and S45F n=1)]. The
remaining 3 tumors contained deletions of 6, 55 and 271 bp, each
including part of exon 3.Reverse transcription-PCR experiments with
RNAs isolated from the adenoma with the 271bp deletion
detectedtranscripts that lacked exon 3, in addition to the normal
transcript. These mutations, as previously reported in severaltypes
of tumors, but never in adrenocortical tumors, abrogate the
phosphorylation-dependent interaction of -cateninknown to lead to
transcription activation of Wnt target genes. Our results suggest,
for the first time, that -catenin genemutations are frequent in
adrenocortical adenomas. Possible involvement of -catenin
activation could contribute toadrenocortical tumorigenesis.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Genome-wide localization of menin and assocated histone
methytransferase complex members. P.C. Scacheri1, S.Davis1, D.T.
Odom2, S. Perkins1, A.M. Spiegel3, P.S. Meltzer1, F.S. Collins1. 1)
NHGRI, Bethesda, MD; 2) WhiteheadInstitute, Cambridge, MA; 3)
NIDCD, Bethesda, MD.
Multiple Endocrine Neoplasia, type I (MEN1) is a familial cancer
syndrome characterized primarily by multipletumors of the endocrine
glands. The gene for MEN1 encodes a ubiquitously expressed tumor
suppressor protein calledmenin. Menin was recently shown to
interact with several components of a trithorax family histone
methyltransferasecomplex including ASH2, Rbbp5, WDR5, and the
leukemia proto-oncoprotein MLL. These findings suggested a rolefor
menin in transcriptional regulation, potentially mediated through
covalent modification of histone H3 at lysine 4. Togain insights
into menin's role as a tumor suppressor, we determined the genomic
occupancy of menin, MLL, andRbbp5, using a strategy that couples
chromatin immunoprecipitation with DNA microarray technology
(ChIP-chip). Themicroarrays used in this approach were custom
designed to contain high density oligonucleotide tiling paths
across20,000 human promoter segments. Given that MLL is reported to
regulate homeobox (HOX) transcription, these arraysalso harbored
tiling paths across all four HOX clusters. Our data in HeLa cells
indicate that menin, MLL, and RBBP5co-localize with RNA polymerase
II to transcriptional start sites, where histone H3 lysine 4
trimethylation occurs.Binding to the HOX clusters is strikingly
different, with broad footprints that extend across inter- and
intragenic regionsof the HOX A and C clusters. To gain insights to
the endocrine-specific tumor phenotype in MEN1, we compared
thesebinding sites to those we mapped in primary human pancreatic
islets, a common site of tumor formation in MEN1patients. Menin was
found to bind to a similar set of target genes in pancreatic
islets, but with notable exceptions at theHOX clusters and
elsewhere. Targets bound by menin exclusively in islet cells
include key transcription factorsinvolved in pancreatic endocrine
development and insulin regulation. We suggest that absence of
menin initiatestumorigenesis by mediating epigenetic changes in
chromatin structure, potentially at the promoters of these
keyendocrine-specific genes.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Expression profiling identifies similar expression pattern of
uveal melanomas with chromosome 3 isodiosmy tothose with monosomy
3. M. Abdel-Rahman1,2, Y. Yang2, F. Davidorf2, C. Eng1. 1) Clinical
Cancer Genetics Prog,Ohio State Univ, Columbus, OH; 2) Department
of Ophthalmology, Ohio Sate Univ, Columbus, OH.
Purpose: Monosomy of chromosome 3 (M3) is the most frequent
somatic alteration in uveal melanomas (UMs). M3has been suggested
as a useful marker for detection of aggressive UMs. Contrary to
earlier reports, recent studiesindicate a high frequency of partial
chromosome 3 deletion in UMs. Moreover, chromosome 3 isodisomy has
beenreported in up to 16% of UMs without M3. The aim of the study
is to identify the expression pattern associated withchromosome 3
isodisomy and partial chromosome 3 alterations compared to M3 in
UMs. Methods: We used acombination of comparative genomic
hybridization (CGH), high resolution genotyping utilizing 38
microsatellitemarkers on chromosome 3 and gene expression profiling
to study 13 primary UMs. The frequency of partialchromosome 3
alterations was studied in a total of 47 UMs. Results: 7/13 UMs
with available CGH data and expressionprofiling showed M3. In
addition 2/13 showed isodiomy of chromosome 3 indicated by minimal
or no alteration ofchromosome 3 by CGH and LOH of most of the
markers on both chromosome 3 arms. The remaining 4/13 UMsshowed
partial gains and/or loss of chromosome 3. All samples with partial
chromosome 3 alterations were confirmedby genotyping. Expression
profiling identified similar expression patterns of samples with
chromosome 3 isodisomy tothose with M3. Expression profiling
identified several regions of common decreased expression on both
chromosomearms 3p and 3q. Finally we identified a high frequency
(22/47) of UMs with partial chromosome 3 alterations.Conclusions:
Our data have revealed similar expression patterns of UMs with
chromosome 3 isodisomy to those withM3, suggesting that isodisomy 3
may portend similar poor prognosis as M3. These results also
indicate that partialchromosome 3 alterations are under-diagnosed
in UMs. The clinical significance of partial chromosome 3
alterationswill need to be further explored. Taken together, our
results are important in designing clinical diagnostic strategies
forthe detection of chromosome 3 alterations in UMs patients.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Advanced genetic analyses on migraine: Trait component analysis
and mouse model expression study. M.Wessman1,2, G. Oswell3, M.
Kallela4, V. Anttila1, M. Kaunisto1, P. Tikka1, E. Hämäläinen1, D.
Nyholt5, J. Terwilliger6,J. Kaprio7, M. Färkkilä4, L. Peltonen8, A.
Palotie1. 1) Finnish Genome Center and Dept of Clin Chem, Univ
Helsinki,Finland; 2) Folkhälsan Research Center, Helsinki, Finland;
3) Dept of Pathology, UCLA, USA; 4) Dept of Neurology,Helsinki
University Central Hospital, Finland; 5) Queensland Inst of Med
Research, Brisbane, Australia; 6) ColumbiaGenome Center, Columbia
University, NYC, USA; 7) Dept of Public Health, Univ Helsinki,
Finland; 8) Dept of HumanMolecular Genetics, National Public Health
Institute, Helsinki, Finland.
Migraine is a common chronic severe headache disorder with a
strong genetic component and a prevalence of ~ 15%in the Caucasian
population. We have collected nearly 700 multiple case families and
based the diagnosis on theInternational Headache Society (IHS)
criteria. This enables us to use medium-to-large size families for
genome-widelinkage analysis and to perform case control association
studies. Our previous genome-wide linkage study identified alocus
on 4q24 linked to the endpoint diagnosis migraine with aura. We
subsequently performed a trait componentlinkage analysis to test
which IHS trait components provide best evidence of linkage to the
4q24 locus and whetheradditional susceptibility regions could be
localized. Strongest evidence of linkage to 4q24 was achieved
withunilaterality (parametric two-point LOD score 4.20),
photophobia (3.73) and phonophobia (3.52). Pulsation was linkedto
17p13 (4.65) and the IHS C-criteria to 18q12 (3.27). To identify
positional candidate genes we performed wholetranscriptome
expression analysis with the Affymetrix 430 2.0 GeneChip on primary
cultures of neurons and glial cellsfrom the tottering mouse with a
mutation in Cacna1a. Using a novel pathway identifying algorithm we
identified 238genes comprising 108 GeneOntology pathways as being
differentially regulated including Ca2+ channel genes Cacna1band
Cacng7, Na+/K+ pumps Atp1a1 and 2, Atp1b1, 2 and 3. We also
identified several genes in our Chr 4 and Chr 17restricted regions.
Candidate genes identified in the mouse expression study are to be
tested in our case control sampleof 1800 subjects.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
A common non-synonymous SNP in the BRI3 (ITM2C) gene accounts
for a significant portion of the linkageevidence for Parkinson
disease on chromosome 2q. N. Pankratz1, L. Byder2, C.A. Halter1, A.
Rudolph3, C.W.Shults4,5, P.M. Conneally1, T. Foroud1, W.C.
Nichols2, Parkinson Study Group-PROGENI Investigators. 1) Medical
&Molec Genetics, Indiana Univ Sch Medicine, Indianapolis, IN;
2) Children's Hospital Research Foundation, Cincinnati,OH; 3) Univ
Rochester, Rochester, NY; 4) Univ California, San Diego, CA; 5) VA
San Diego Healthcare System, SanDiego, CA.
Previously, we identified significant linkage to chromosome 2q35
using a sample of individuals with familialParkinson disease (PD).
The linkage was strongest among those families with verified PD and
those with a strongerfamily history (four or more affected or an
affected sibpair with an affected parent). The disease allele
appeared to beinherited in an autosomal dominant fashion, and
penetrance appeared to be high in those individuals with a
strongfamily history. Subsequently, we have genotyped six
additional markers in the linkage region. With these
additionalmarkers, the autosomal dominant LOD score for the subset
remained significant, and the non-parametric linkageevidence of the
verified PD sample increased from a LOD score of 2.5 to 4.0. One
candidate gene in this region is BRI3(ITM2C), which belongs to a
family of integral membrane proteins. One of its homologs (BRI2) is
associated withfamilial dementia. Little is known about BRI3,
except that it is highly enriched in human brain and that it
interacts withBACE1, an enzyme responsible for converting
beta-amyloid, a protein involved in Alzheimer disease
pathogenesis.When we screened our PD sample for mutations in the
BRI3 gene, we identified a rare non-synonymous point
mutation(Arg236Trp) in one of our patients, as well as a
non-synonymous polymorphism (Gly53Ser) that appears to be commonin
the general population (allele frequency = 0.40). Using the
Genotype-IBD Sharing Test (GIST) under the samehypothesis of
autosomal dominant inheritance, we have shown that this common SNP
accounts for a significant portionof the linkage to chromosome 2q
(p=0.03). It is therefore possible that this polymorphism is a
common PDsusceptibility gene with low penetrance (similar to APOE
for Alzheimer disease).
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
High-density linkage screen in multiple sclerosis. J. Haines,
for the International Multiple Sclerosis GeneticsConsortium. Ctr
Human Genetics, Vanderbilt Univ Medical Ctr, Nashville, TN.
Ten centimorgan microsatellite map have been the standard tool
used for whole genome linkage screening since themid 1990s and to
date 11 screens employing this methodology have been published in
multiple sclerosis. Inspection ofthese data shows that average
genotyping success and information extraction rates are
disappointing (80% and 44%respectively). Moreover available
evidence suggests that genotyping errors rates in these studies
(particularly the olderones) are likely to be > 1% for many if
not most of the markers typed. The advent of high throughput cost
effective SNPgenotyping has provided new whole genome linkage
screening tools promising significantly improved genotypingsuccess
rates and accuracy, which at high density should substantially
enhance information extraction. Prompted by thepotential benefits
of these new tools we have typed the Illumina BeadArray linkage
mapping panel (version 4) in a setof 730 multiplex families from
Australia (97), Scandinavia (165), the United Kingdom (298) and the
United States(170), which together provide 945 affected relative
pairs (830 sib pairs, 14 half sib pairs, 54 cousin pairs and
47avunclular pairs). After excluding 474 markers with genotyping
success rates of 0.16) data is available from 4506 markers spanning
the genome.A total of 2709 samples have been typed. The mean
information extraction is 79.3% (range 42.4 - 91.3%) and
theobserved Mendelian inconsistencies suggest that within this data
the genotyping error rate is just 0.002%. Highlysignificant linkage
is observed in the region of the MHC (MLS 11.7) while suggestive
linkage is seen on chromosomes17q23 and 5q33. This screen provides
the most definitive linkage map for multiple sclerosis currently
available andillustrates the substantial increase in power that can
be achieved using high density SNP linkage mapping sets. It is
clearthat future attempts to identify non-MHC susceptibility genes
in multiple sclerosis will have to involve large samplesizes and
association based methodology.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Homozygosity mapping with high-density SNP arrays identifies a
novel gene for Bardet-Biedl Syndrome. A.P.Chiang1, J.S. Beck1,2,
A.L. Ferguson1, K. Elbedour3, R. Carmi3, H.-J. Yen1,2, M.K.
Tayeh1,2, J. Wei1,2, D.Y.Nishimura1, T.A. Braun1, E.M. Stone1,2,
T.L. Casavant1, V.C. Sheffield1,2. 1) University Iowa, Iowa City,
IA, USA; 2)Howard Hughes Medical Institute, Iowa City, IA, USA; 3)
Genetic Institute, Ben Gurion University of the Negev, Beer-Sheva,
Israel.
The identification of mutations in genes that cause human
diseases has largely been accomplished through the use ofpositional
cloning, which relies on linkage mapping. In studies of rare
diseases, the resolution of linkage mapping islimited by the number
of meioses and marker density. One recent technological advance is
the development of highdensity Single Nucleotide Polymorphism (SNP)
microarrays for genotyping. The SNP arrays overcome low
markerinformativity by using a large number of markers to achieve
greater coverage at finer resolution. To determine the utilityof
this technology for homozygosity mapping using small pedigrees, we
genotyped a small consanguineous IsraeliBedouin (AT) family with
autosomal recessive Bardet-Biedl Syndrome (BBS; obesity, pigmentary
retinopathy,polydactyly, hypognoadism, renal and cardiac
abnormalities, and cognitive impairment). While eight genes have
beenidentified to cause BBS, over half of BBS patients have unknown
genetic defects, leaving the possibility of additionalgenes, that
when mutated, cause BBS. The AT family was a good candidate for
homozygosity mapping with highdensity SNP arrays because it was not
linked to any of the known BBS loci and a genome-wide scan with
microsatellitemarkers at ~10 cM density did not reveal a linked
locus. DNA from all four individuals affected with BBS from
thefamily was hybridized to the HindIII chip of the Affymetrix 100K
SNP chip set and the resulting data analyzed forregions of
homozygosity. Regions of homozygosity were prioritized based on the
physical distance, the number ofconsecutive homozygous SNPs, and
marker density. Mutation analysis in the best candidate
homozygosity regionidentified a gene with a conserved homozygous
missense mutation. Functional analysis of this gene in a
zebrafishsystem provides additional evidence that this is a BBS
gene (BBS9).
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Genomeutwin-European twin cohorts: Combined analysis of genome
scans guide to two QTL:s for body statureincluding one common
Caucasoid X-chromosomal locus. L. Peltonen1, S. Sammalisto1, T.
Hiekkalinna1, N. Martin2,J. Harris3, D. Boomsma4, K. Christensen5,
K. Ohm Kyvik5, N. Pedersen6, T. Andrew7, T. Spector7, E. Widen8,
A.Palotie8, M. Perola1, GenomEUtwin. 1) Dept Med Genetics &
Molec Med, Univ Helsinki & Nat Pub Hth In, Helsinki,Finland; 2)
Queensland Institute of Medical Research, Brisbane, Australia; 3)
Norwegian Institute of Public Health,Oslo; 4) Free University,
Amsterdam, Netherlands; 5) University of Southern Denmark, Odense,
Denmark; 6)Karolinska Institutet, Stockholm, Sweden; 7) St Thomas
Hospital, London, United Kingdom; 8) University of Helsinki,Finnish
Genome Center, Finland.
Twin cohorts provide a unique advantage for investigations of
the role of genetics and environment in the etiology ofcommon
traits. Co-twins share environment throughout critical fetal period
and early years of life and twin designharmonizes this component of
complex traits in a unique manner. The EU-funded
GenomEUtwin(www.genomeutwin.org) consortium consists of eight twin
cohorts (Australian, Danish, Dutch, Finnish, Italian,Norwegian, UK
and Swedish) with the total number of 600000 twin pairs. Federated
database with open source codehas been created to share the data
across the cohorts. We performed QTL analysis of stature (body
height) usinggenome-wide scans performed for 8775 individuals:
Australia (n=3730), Denmark (628), Finland (772),
Netherlands(1313), Sweden (102) and United Kingdom (2230). The
marker maps were combined and related to the sequencepositions
using computer program developed by us which uses DeCode genetic
map markers as an anchoring
set(www.bioinfo.helsinki.fi/cartographer). We used the program
Merlin for variance components analysis with age, sexand
country-of-origin as covariates. The covariate adjusted
heritability was 82% for stature in the pooled data set. Wefound
evidence for two major QTLs for human stature on 15q24 (LOD=3.75,
1-lod drop 11cm) and Xq25 (LOD=2.73,15cM). Especially the evidence
for linkage for the X-chromosomal locus is contributed by most of
the cohorts, thussuggesting an evolutionally old genetic variant
having effect on the growth in European-based populations.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Evidence for association between hepatic nuclear factor 4, alpha
and triglycerides in Finnish familial combinedhyperlipidemia
families. D. Weissglas1, E. Suviolahti1,2, J. Lee1, A. Jokiaho1,
M-R. Taskinen3, P. Pajukanta1. 1)Department of Human Genetics,
UCLA, Los Angeles, CA; 2) Department of Molecular Medicine,
National PublicHealth Institute, Biomedicum, Helsinki, Finland; 3)
Department of Medicine, University of Helsinki, Finland.
Familial combined hyperlipidemia (FCHL), characterized by
increased levels of serum total cholesterol, triglycerides(TGs) or
both, is observed in up to 20% of patients with premature coronary
heart disease. Previously we identified alocus linked to elevated
TGs on chromosome 20q13.11 in Finnish FCHL families. Numerous
linkage studies of type 2diabetes mellitus (T2DM) have also found
evidence of linkage to this region. Recently, several independent
groupsidentified associations between single nucleotide
polymorphisms (SNPs) in the hepatic nuclear factor 4, alpha
(HNF4A)gene on 20q12-q13.1 and T2DM in Finnish and Ashkenazi Jews.
Because there is a clear phenotypic overlap betweenFCHL and T2DM,
we tested this gene region for association with FCHL in Finnish
families. To date no associationbetween plasma lipid levels and
HNF4A has been identified in FCHL. We constructed the linkage
disequilibrium (LD)structure of the HNF4A region by using both the
HapMap data and recently published Finnish LD data. Based on
thatinformation, we genotyped the seven SNPs, previously associated
with T2DM, and three additional tag-SNPs fromdistinct haplotype
blocks to capture most of the genetic variation within the HNF4A
gene, including promoters 1 and 2.A SNP in promoter 2 of the HNF4A
gene was significantly associated with high TGs (P=0.007) in 720
family membersof 60 FCHL families. Haplotype analyses also showed
significant associations with TGs as well as with
plasmaapolipoprotein B and high-density lipoprotein levels for
several HNF4A haplotypes. The most significant haplotype forhigh
TGs (P=0.008) has a frequency of 0.289 in the Finnish FCHL
families. In conclusion, this study is the first toprovide
significant evidence for association between the intragenic
variants of the HNF4A gene and plasma lipid levelsin FCHL.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
A new method to correct for population stratification in genetic
case-control association studies: centralizing thenon-central
chi-square. P. Gorroochurn1, G.A. Heiman2, S.E. Hodge1,3, D.A.
Greenberg1,3. 1) Division of StatisticalGenetics,Dept
Biostatistics, Columbia Univ, New York, NY; 2) Department of
Epidemilogy, Columbia Univ, NewYork, NY; 3) Clinical-Genetic
Epidemiology Unit, New York State Psychiatric Institute, New York,
NY.
Recently, several authors have investigated the use of genomic
information in an attempt to eliminate bias due topopulation
stratification (PS) in case-control association studies. We here
present a new method, the delta-centralization(DC) method, to
correct for PS. DC works well even when there is a lot of bias due
to PS. In the presence of PS, theusual chi-square statistics used
to test for association have non-central chi-square distributions.
Other methods approachthe non-centrality indirectly, but we deal
with it directly, by estimating the non-centrality parameter
itself. Specifically:(1) We define a quantity delta, a function of
the relevant subpopulation parameters, that exactly predicts (in
relativelylarge samples) the elevation of the false positive rate
due to PS, when there is no true association between markergenotype
and disease. (Delta is quite different from F_ST and can be large
even when F_ST is small.) (2) We show howto estimate delta, using a
panel of unlinked neutral loci. (3) We show that the square of
delta corresponds to the non-centrality parameter of the chi-square
distribution. Thus, we can centralize the chi-square using our
estimate of delta;this is the DC method. (4) We demonstrate, via
computer simulations, that DC works well with as few as
25-30unlinked markers, where the markers are chosen to have allele
frequencies reasonably close (within 0.1) to those at thetest
locus. (5) We compare DC with genomic control and show that whereas
the latter becomes over-conservative whenthere is considerable bias
due to PS (i.e. when delta is large), DC performs well for all
values of delta.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
A Spatial Clustering Approach to the Analysis of Genetic
Association Studies. E.R.B. Waldron1, J.C. Whittaker1,2,D.J.
Balding1. 1) Epidemiology and Public Health, Imperial College
London, London, United Kingdom; 2)Epidemiology and Population
Health, London School of Hygiene and Tropical Medicine, London.
Haplotype-based approaches to the analysis of genetic
association studies have important advantages over single-marker
analyses, but they nevertheless suffer from limitations. For
instance, rare haplotypes may not analysedeffectively. Furthermore,
the relationships between distinct but similar haplotypes that may
share recent commonancestry are often not accounted for. We have
developed an algorithm based on spatial statistics techniques,
employinga simple distance in haplotype space that reflects
evolutionary processes. The algorithm searches for case-rich
clustersof similar haplotypes. Membership of this cluster
corresponds to predicting the allele at an unobserved causal SNP.
Thealgorithm can be applied to fine-mapping if the distance metric
depends on the putative location of a causal allele. Usingthis
algorithm to analyse data for the CYP2D6 gene, for which the true
causal polymorphism is fully characterised, wecorrectly predicted
nearly 98% of the genotypes at the major causal polymorphism and
the functional mutation wasmapped accurately. Simulation studies
also revealed consistently better performance than alternative
fine-mappingalgorithms and allowed us to identify situations in
which multi-point approaches offer a substantial improvement
oversingle-point analyses.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Admixture-Matched Case-Control Study: A Practical Approach for
Genetic Association Studies in AdmixedPopulations. J. Kho1,2, H.J.
Tsai1,2, N. Shaikh1,2, S. Choudry1,2, M. Navqi1,2, D. Navarro1,2,
H. Matallana1,2, R.Castro1,2, C. Lilly3, H.G. Watson4, K. Meade5,
M.L. Noir6, S. Thyne1, E. Ziv1, E.G. Burchard1,2. 1) UCSF,
SanFrancisco, CA; 2) Lung Biology Center, San Francisco General
Hospital, SF, CA; 3) Brigham and Womens Hospital,Boston, MA; 4) The
James A. Watson Wellness Center, Oakland, CA; 5) Children's
Hospital and Research Institute,Oakland, CA; 6) Bay Area
Pediatrics, Oakland, CA.
Case-control genetic association studies in admixed populations
are known to be susceptible to genetic confoundingdue to population
stratification. The transmission/disequilibrium test (TDT) approach
can avoid this problem. But, theTDT is expensive and impractical
for late-onset diseases. Case-control study designs, in which cases
and controls arematched by admixture, can be an appealing
alternative for genetic association studies in admixed populations.
Weapplied this matching strategy when recruiting our African
American participants in the Study of African American,Asthma,
Genes and Environments (SAGE). Group admixture in this cohort
consists of 83% African ancestry and 17%European ancestry. By
carrying out several complementary analyses, our results show that
there is substructure, but thatthe admixture distributions are
almost identical in cases and controls, and also in cases only. We
performed associationtests for asthma-related traits with ancestry,
and only found that FEV1, a measure of asthma severity, was
associatedwith ancestry after adjusting for socio-economic and
environmental factors (p=0.01). We did not observe an excess oftype
I error rate in our association tests for ancestry informative
markers and asthma-related traits when ancestry wasnot adjusted in
the analyses. Furthermore, using the association tests between
genetic variants in a known asthmacandidate gene, 2 adrenergic
receptor (2AR) and FEF25-75, a measure of bronchodilator drug
responsiveness to theasthma medication, albuterol, as an example,
we showed population stratification was not a confounder in our
study.Our work demonstrates that admixture-matched case-control
strategies can efficiently control for populationstratification
confounding in admixed populations.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Analysis of fine structure recombination patterns in a human
chromosomal region not previously known toharbor a recombination
hot spot. I. Tiemann-Boege1, P. Calabrese1, D.M. Cochran1, R.
Sokol2, N. Arnheim1. 1)Molecular and Computational Biology Program,
University of Southern California, Los Angeles, CA; 2) Health
ScienceCenter, University of Southern California, Los Angeles,
CA.
For decades, evidence has suggested that crossing over is not
homogenous across the human genome but can behighly localized in
hot spots. Recent molecular studies of regions known to contain a
hot spot based on classicalmethods have revealed that hot spots can
be present where strong linkage disequilibrium (LD) is interrupted
by patternsof very low LD. Whether recombination is generally
restricted to hot-spots and whether hot spots can be found in
largerchromosomal regions without exceptional frequencies of
crossing over is yet to be addressed. Using sperm typing
thatselectively amplifies only recombinants in a pool of sperm
genomes, we analyzed the recombination fractions in smallcontiguous
intervals (~5 kb) of a 104 kb region of human chromosome 21 that
showed no exceptional crossoveractivity. The observed recombination
intensity ranges from 0.10 (95% confidence interval: 0.01, 0.29) to
12.21 (9.67,15.05) times the human genome average rate. We
identified two hot-spots with a recombination activity of 11.17
(9.98,12.42) and 12.21 (9.67, 15.05) times the human average and
four regions with a frequency ranging between 0.93 (0.59,1.33) and
3.62 (3.09, 4.19) fold human average. The remaining regions all
have an estimated frequency below 0.55times the human average (all
have a 95% upper bound below 0.84). The recombination patterns
observed by spermtyping were compared with the patterns predicted
by LDHat, Hotspotter and a third algorithm developed by one of
us(P.C.). Patterns are roughly congruent with the recombination
inferred from the population estimators with someexceptions. Our
data suggests that it is likely that hotspots are usually found in
larger chromosomal regions withoutexceptional frequencies of
crossing over. Clearly, studies on other chromosomal intervals are
needed to get a bettergenome-wide picture of the extent to which
recombination is generally restricted to hot-spots and the degree
to whichLD analyses can predict recombination intensities.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Improved Association Analyses of Disease Subtypes in Case-Parent
Triads. M.P. Epstein1, I.D. Waldman2, G.A.Satten3. 1) Dept Human
Genetics, Emory Univ, Atlanta, GA; 2) Dept Psychology, Emory Univ,
Atlanta, GA; 3) Centersfor Disease Control and Prevention, Atlanta,
GA.
The sampling of case-parent triads is an appealing strategy for
conducting association analyses of complex diseases.In certain
situations, one may have interest in using the triads to identify
genetic variants that are associated with aspecific subtype of
disease, perhaps related to severity of symptoms or sensitivity to
medication. A straightforwardstrategy for conducting such a subtype
analysis would be to analyze only those triads with the subtype of
interest. Whilesuch a strategy is valid, we show that triads
without the subtype provide additional genetic information that
increasespower to detect association with the subtype. We
incorporate this additional information using a
likelihood-basedframework that permits flexible modeling and
estimation of allelic effects on disease subtypes and also allows
formissing parental data. Using simulated data under a variety of
genetic models, we show that our proposed associationtest
consistently outperforms association tests that only analyze
subtype triads. We also apply our method to a triadstudy of
attention-deficit hyperactivity disorder and identify a genetic
variant in the dopamine transporter gene that isassociated with a
hyperactive-impulsive subtype.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Estimating genetic effects from genomewide association scan
data. S. Zollner, J. Pritchard. Dept Human Genetics,Univ Chicago,
Chicago, IL.
Genomewide association studies are a powerful method to detect
loci harboring variants that affect complex traits.After a
successful scan for association it is of great interest to estimate
the impact of a detected variant on the phenotypeof a carrier
(penetrance) and on the population as a whole (frequency). These
estimates allow to assess the importanceof a mutation, they may
provide information about its biological effect and facilitate
planning replication studies. Asthese estimations are usually
performed based on the same data that were used to identify a
variant, the results areaffected by ascertainment bias, causing the
genetic effects of a variant to be overestimated if not corrected
for. Thisoverestimation is considered the main reason that many
replication studies fail as the sample size needed
isunderestimated. The magnitude of this bias depends on the power
of the initial association study. Here, we present anapproach that
corrects for the ascertainment effect and generates an approximate
maximum-likelihood estimate of thefrequency of a variant and its
penetrance parameters. The method produces both a point-estimate
and a confidenceregion for all parameters. We apply this estimator
to simulated datasets and study its performance and its ability
todistinguish between different models of penetrance. We show that
by taking epidemiological data into account, it isoften possible to
obtain fairly precise estimates of all parameters, even if the
power of the genome scan is low. We alsoshow that the error of the
estimate decreases with sample size, independent of the power of
the original test forassociation. Finally, we demonstrate that the
same algorithm can be used to assess interactions between a genetic
variantand environmental risk factors.
Copyright © 2005 The American Society of Human Genetics. All
rights reserved.
-
Regression-Based QTL analysis method incorporating
parent-of-origin effect. O. Gorlova, L. Lei, D. Zhu, Y.Zhang, W.
Li, S. Shete, R.A. Price, C. Amos. Dept Epidemiology, Univ Texas MD
Anderson CA Ctr, Houston, TX.
We present an extension of Sham et al.s (2002) regression-based
quantitative-trait linkage analysis method toincorporate
parent-of-origin effects. We separately regressed total, paternal,
and maternal IBD sharing on traits squaredsums and differences. We
also developed a test for imprinting that indicates whether there
is any difference betweenpaternal and maternal regression. We use a
panel of statistics to detect imprinting, which includes an overall
T statistic(a test for total linkage), both parental T statistics
(tests for parental-specific linkages), and the D index (a test
forimprinting). We performed an extensive simulation to examine the
performance of the panel. We found that when usingempirical
percentiles the method is very powerful in detecting
parent-specific linkage with correct type I error rate forthe
non-linked parental component. Missing parental genotypes increase
the type I error rate of both the linkage andimprinting tests and
decreases the power of the imprinting test. When the major gene has
low heritability, the power ofthe method decreases dramatically but
the panel still performs well. We also used a permutation
algorithm, which canensure the appropriate type I error rate. We
applied the method to a data from a study of 6 body size related
measuresand 23 loci on chromosome 7 for 255 nuclear families.
Multivariate identities-by-descent were obtain using amodification
of SIMWALK program. A parent-of-origin effect consistent with
maternal imprinting was suggested at99.67-111.26 Mb for body mass
index, bioelectrical impedance analysis, waist circumference, and
leptin concentration.
Copyright © 2005 The American Soci