Development, Implementation and Testing of a DNA Microarray Test Suite Ehsanul Haque Mentors: Dr. Cecilie Boysen Dr. Jim Breaux ViaLogy Corp.

Development, Implementation Development, Implementation and Testing of a DNA and Testing of a DNA

Microarray Test SuiteMicroarray Test Suite

Ehsanul HaqueEhsanul Haque

Mentors:Mentors:Dr. Cecilie BoysenDr. Cecilie Boysen

Dr. Jim BreauxDr. Jim BreauxViaLogy Corp.ViaLogy Corp.

OutlineOutline

PurposePurpose

BackgroundBackground

DatasetsDatasets

AlgorithmsAlgorithms

ResultsResults

ConclusionConclusion

PurposePurpose

To design tests to assess the results fromTo design tests to assess the results from

Different microarray data analysis servicesDifferent microarray data analysis services VMAxS, GCOSVMAxS, GCOS

Different updates of the same data analysis serviceDifferent updates of the same data analysis service VMAxS_V1, VMAxS_V2, VMAxS_V3, VMAxS_V4VMAxS_V1, VMAxS_V2, VMAxS_V3, VMAxS_V4

Different background correction, normalization and Different background correction, normalization and summarization methodssummarization methods

VMAxS, PLIER, RMA, dCHIPVMAxS, PLIER, RMA, dCHIP Different technologies of microarray data analysisDifferent technologies of microarray data analysis

Affymetrix, ABI, Agilent, GE Healthcare, IlluminaAffymetrix, ABI, Agilent, GE Healthcare, Illumina

Dataset processed byDataset processed by

DEGDEG

MoreMore

RelRel. . AccuAccu..CVCV

Method DMethod DMethod CMethod CMethod BMethod BMethod AMethod A

TestsTests Dataset processed byDataset processed by

DEGDEG

MoreMore

RelRel. . AccuAccu..CVCV

Method DMethod DMethod CMethod CMethod BMethod BMethod AMethod A

TestsTests

BackgroundBackground

VMAxS RA Report

(Resonance Amplitude)

Probe level signal data

11-20 probe per gene

1 expression value per gene

(if from Affymetrix system)

VMAxSSignal Detection via

QRI

DAT or raw image fileOnly raw data files accepted

Quantum Resonance Interferometry (QRI)Quantum Resonance Interferometry (QRI)– – Active Signal ProcessingActive Signal Processing

From image to gene level dataFrom image to gene level data

Raw pixel level data

VMAxS/GCOS

Probe level data

VMAxS/MAS5.0/PLIER/RMA

Background correction

Normalization

Summarization (PM/MM)

Gene level data

Image analysis

DatasetsDatasets

.DAT File .DAT File

Method D

Method C

Method B

Method A

Each Dataset stems from the same raw Each Dataset stems from the same raw AffymetrixGeneChip .dat filesAffymetrixGeneChip .dat files

GeneChip family Human 2A: GeneChip family Human 2A: 54676 54676 genesgenes

3 Replicates * 8 Dilutions

3 Replicates * 8 Dilutions

3 Replicates * 8 Dilutions

3 Replicates * 8 Dilutions

24 Samples

24 Samples

24 Samples

24 Samples

Test SuiteTest Suite [R scripting][R scripting]

Master Function

Function CV(Coefficient of Variance)

User File User File

Function RA(Relative Accuracy)

Function DEG(Differential Expressed Gene)

OtherOutput

File list & parameters

Algorithm Algorithm

Function RA(Rela. Accu.)

Function TABLE.IN

Using a loop, calls function TABLE.IN to work on the User File list of input datasets

Calls other functions, parses different formats of data-sets and provides function RA with the required table.

Method A Method B Method C Method D

Slope 0.497897086864 0.497897086864 0.394078146688 0.394078146688 0.33660070237 0.33660070237 0.2112750562570.211275056257

User File User File File list & parameters

Calculates Mean & SD

Calculates Slope and Ave. of mean intensities

Plots Slope vs. Average of mean intensity

PlotsConc. vs. Mean Intensity

Creates a list ofMedian Slope

Replaces zero values

CV PlotCV Plot Three Replicates A, B Three Replicates A, B & C& C

CV Plot CV Plot With fitted lineWith fitted line

8 Dilutions8 Dilutions Mean Intensity vs. Mean Intensity vs. ConcentrationConcentration

Slope vs. Mean IntensitySlope vs. Mean Intensity

Median SlopeMedian Slope

ResultsResults

TestsTestsDataset processed byDataset processed by

Method Method AA

Method Method BB

Method Method C C

Method Method DD

CVCV 0.117055480.1170554806340634

0.1719023230.171902323603603

0.0727114620.07271146285468546

0.1008522740.100852274894894

Rel. Rel. Accu.Accu.

0.497897080.4978970868646864

0.3940781460.394078146688688

0.3366007020.3366007023737

0.2112750560.211275056257257

DEGDEG

MoreMore

ConclusionConclusion Developed and implemented two tests in RDeveloped and implemented two tests in R

CV & RA (relative accuracy)CV & RA (relative accuracy)

Tested 2 X 4 datasetsTested 2 X 4 datasets Different parameters of the same analysis Different parameters of the same analysis

methodmethod Different normalization methodsDifferent normalization methods

More tests are needed to draw a better More tests are needed to draw a better conclusion about the data. conclusion about the data.

Reference & Reference & AcknowledgmentAcknowledgment

References:References:

Zhijin Wu el al, Zhijin Wu el al, A Model Based Background Adjustment of A Model Based Background Adjustment of Oligonucleotide Expression ArraysOligonucleotide Expression Arrays, Johns Hopkins University, Dept , Johns Hopkins University, Dept of Biostatistics Working Papers, 2004, paper 1 of Biostatistics Working Papers, 2004, paper 1

Irizarry et al, Irizarry et al, Exploration, Normalization, and Summaries of High Exploration, Normalization, and Summaries of High Density Oligonucleotide Array Probe Level DataDensity Oligonucleotide Array Probe Level Data, Biostatistics (2003), , Biostatistics (2003), 4, 2, pp. 249-2644, 2, pp. 249-264

Cope et al, Cope et al, A Benchmark of Affymetrix GeneChip Expression A Benchmark of Affymetrix GeneChip Expression Measures, BioinformaticsMeasures, Bioinformatics, Vol. 1 no. 1 2003, pp. 1-10, Vol. 1 no. 1 2003, pp. 1-10

Acknowledgements:Acknowledgements:

Dr. Cecilie BoysenDr. Cecilie Boysen Dr. Jim BreauxDr. Jim Breaux

- - VialogyVialogy

Southern California Bioinformatics Summer Institute (SoCalBSI)Southern California Bioinformatics Summer Institute (SoCalBSI) National Institute of Health (NIH)National Institute of Health (NIH)