A Comprehensive Analytical Pipeline for a Genome-Wide Association Study of Bronchopulmonary Dysplasia: From SNP to Copy Number Variation and from Gene to Pathway

• Activity Number: 349
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Hui Wang*+ and Krystal R St. Julien and David K Stevenson and Thomas J. Hoffmann and John S Witte and Laura C. Lazzeroni and Mark A. Krasnow and Cele C. Quaintance and John W. Oehlert and Laura L. Jelliffe-Pawlowski and Jeffrey B. Gould and Gary M Shaw and Hugh O'Brodovich
• Companies: Stanford University and Stanford School of Medicine and Stanford School of Medicine and University of California San Francisco and University of California San Francisco and Stanford University and Stanford School of Medicine and Stanford School of Medicine and Stanford School of Medicine and California Genetic Disease Screening Program and Stanford School of Medicine and Stanford School of Medicine and Stanford School of Medicine
• Keywords: GWAS ; association testing ; high dimensional data ; genetic marker evaluation

Abstract

Bronchopulmonary dysplasia is a major lung morbidity in premature infants. Twin studies indicate that it is a heritable condition with an estimated heritability as high as 80%. We conducted a genome-wide association study on ~1,700 premature infants born in California between 2005 - 2008 using an Illumina 2.5M platform. Source DNA was derived from newborn blood spots stored by the State of California. This presentation will focus on unique data challenges that can arise in conducting such genetic inquiries in pediatric populations as well as the development of a comprehensive analytic pipeline for such data. We will discuss the efficacy of available tools including PLINK and several machine learning algorithms and the strategies to efficiently analyze both SNP and copy number variations derived from genotyping intensities. We found that, in terms of ranking variants, considering both p-values and odds ratios performed better than considering either. We will also discuss the pipeline and strategy to impute and analyze GWAS data in highly heterogeneous California population.