Evaluating Rare Variants in Genomewide Studies

• Activity Number: 646
• Type: Invited
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): John S. Witte*+
• Companies: University of California, San Francisco
• Keywords: sparse data ; genetics ; rare variants ; empirical Bayes ; genome wide association studies

Abstract

Genome-wide studies have successfully detected numerous highly replicated associations between common variants and disease. But they explain little of disease heritability, so it is also important to study rare variants. Due to the low frequency of such variants, however, the analysis of their potential association with disease is problematic. A few analytic approaches have been used, ranging from treating each rare variant independently and obtaining an enormous sample size to collapsing rare variants into a single 'marker' to increase power. Here we propose compromising between these extremes with an approach that evaluates each rare variant individually, but borrows strength from other variants that have similar biological properties. We motivate the approach with an application to rare variants in studies of birth defects.