Quantifying Missing Heritability and Estimating Genetic Effects for Complex Traits Due to Rare Variants: Application to Sequence-Based Association Studies

• Activity Number: 455
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Suzanne Leal*+ and Dajiang Liu*+
• Companies: Baylor College of Medicine and William Marsh Rice University
• Keywords: association analysis ; rare variants ; genetic variance ; complex traits ; heritability

Abstract

Next-generation sequencing has led to a surge of complex-trait rare variant (RV) association studies. Detecting associations with individual RVs is grossly underpowered, therefore many association tests have been developed, which aggregate RVs across a genetic region, e.g. gene. When multiple variants are jointly analyzed, it is impossible to tease apart causal from non-causal variants. Therefore, it is only possible estimate genetic average effects (GAE) for a group of variants, which is defined by the mean quantitative trait differences between variant carriers and non-carriers. The estimates of GAE can be grossly inflated due to the winner's curse. We extended the bootstrap-split-sample (BSS) algorithm for RV analyses and demonstrate that the BSS corrected estimates are accurate. Due to the presence of non-causal variants or causal variants with different magnitudes and directions, the GAE based genetic variance is no greater than the true locus genetic variance, and the BSS corrected estimates are the lower bound for the true locus genetic variance. Even if RVs play a major role in the etiology of complex traits, because of aggregate analysis, the genetic variance can be under