After more than a decade of improvements to experimental and data analysis techniques, microarray technology is poised to become instrumental in the era of personalized genomics. In fact, Affymetrix, a leading manufacturer, recently achieved the first FDA clearance of high-throughput gene profiling reagents. Microarrays were also crucial in the successful development of an FDA approved breast cancer recurrence assay - making it possible to identify patients at risk of distant recurrence following surgery. Moreover, approximately one half of all Pub Med publications citing microarrays were published during the last two years. We therefore expect laboratories in academia and industry to continue relying on these technologies for several years as newer genomic technologies mature, and that manufacturers will continue to develop new products at a rapid pace. All microarray data analyses begin by converting raw measures into the data and summary statistics relied upon by biologists and clinicians. This first step, referred to as preprocessing, has an enormous influence on the quality of the ultimate measurements and results from studies that rely upon them. Our group has previously demonstrated that statistical methodology can provide great improvements over ad hoc data analysis algorithms offered as defaults by array manufacturers. Our highly cited statistical methodology and our widely used software implementations demonstrate the success of our work. While gene expression has been the most popular microarray application, recently, the technology has been used to measure diverse genomic endpoints including genotype, copy number variants, transcription factor binding sites, and several epigenetic marks, including DNA methylation. During the first funding period, our group was dedicated to understanding the bias and systematic errors which can obscure results, thwart discovery, and contribute to findings that are not reproducible. We have amassed expertise and developed successful data analysis tools to effectively preprocess raw data, making the technology prime for translational research and clinical applications. However, this transition from basic to clinical research will generate new statistical challenges and our methodology, which has partly facilitated the success of microarrays, will play an important role in the promising next period of research driven by microarray technology. Our goal is to develop the next generation of preprocessing and analysis tools with an emphasis on translational applications. Toward this goal, the current proposal has the following specific aims: developing single array preprocessing methodology with emphasis on batch effect removal, developing microarray analysis tools for three urgent needs, and developing generalized bump hunting methodology for detecting differentially methylated regions.

Public Health Relevance

Microarrays are poised to become instrumental in the era of personalized genomic. The first step in microarray data analyses convert raw data into the summaries relied upon by biologists. Our group will develop statistical methodology for the next generation of microarray applications.

Agency
National Institute of Health (NIH)
Institute
National Institute of General Medical Sciences (NIGMS)
Type
Research Project (R01)
Project #
2R01GM083084-06
Application #
8297156
Study Section
Biostatistical Methods and Research Design Study Section (BMRD)
Program Officer
Sheeley, Douglas
Project Start
2007-09-24
Project End
2016-08-31
Budget Start
2012-09-01
Budget End
2013-08-31
Support Year
6
Fiscal Year
2012
Total Cost
$318,559
Indirect Cost
$116,519
Name
Johns Hopkins University
Department
Biostatistics & Other Math Sci
Type
Schools of Public Health
DUNS #
001910777
City
Baltimore
State
MD
Country
United States
Zip Code
21218
Kang, Jian; Bowman, F DuBois; Mayberg, Helen et al. (2016) A depression network of functionally connected regions discovered via multi-attribute canonical correlation graphs. Neuroimage 141:431-41
Vandiver, Amy R; Irizarry, Rafael A; Hansen, Kasper D et al. (2015) Age and sun exposure-related widespread genomic blocks of hypomethylation in nonmalignant skin. Genome Biol 16:80
De Livera, Alysha M; Sysi-Aho, Marko; Jacob, Laurent et al. (2015) Statistical methods for handling unwanted variation in metabolomics data. Anal Chem 87:3606-15
Hicks, Stephanie C; Irizarry, Rafael A (2015) quantro: a data-driven approach to guide the choice of an appropriate normalization method. Genome Biol 16:117
Maksimovic, Jovana; Gagnon-Bartsch, Johann A; Speed, Terence P et al. (2015) Removing unwanted variation in a differential methylation analysis of Illumina HumanMethylation450 array data. Nucleic Acids Res 43:e106
Liu, Ruijie; Dai, Zhiyin; Yeager, Meredith et al. (2014) KRLMM: an adaptive genotype calling method for common and low frequency variants. BMC Bioinformatics 15:158
Bureau, Alexandre; Parker, Margaret M; Ruczinski, Ingo et al. (2014) Whole exome sequencing of distant relatives in multiplex families implicates rare variants in candidate genes for oral clefts. Genetics 197:1039-44
Aryee, Martin J; Jaffe, Andrew E; Corrada-Bravo, Hector et al. (2014) Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics 30:1363-9
Parker, Hilary S; Corrada Bravo, Héctor; Leek, Jeffrey T (2014) Removing batch effects for prediction problems with frozen surrogate variable analysis. PeerJ 2:e561
Bureau, Alexandre; Younkin, Samuel G; Parker, Margaret M et al. (2014) Inferring rare disease risk variants based on exact probabilities of sharing by multiple affected relatives. Bioinformatics 30:2189-96

Showing the most recent 10 out of 87 publications