The Mystery of Missing Heritability: The Role of Rare Variants

• Activity Number: 659
• Type: Invited
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Or Zuk*+
• Companies: The University of Chicago
• Keywords: missing heritability ; common variants ; rare variants ; exome studies ; Wright-Fisher model ; Two-class model

Abstract

Genetic association studies performed in recent years are biased towards discovering common variants, due to technological limitations and statistical power issues. Multiple yet-to-be discovered rare variants with large effects have been proposed as a major source of missing heritability in humans, but no systematic analysis of their contribution to heritability has been performed, mainly due to lack of adequate data.

Testing for associations for rare variants and assessing their contribution to heritability has important clinical and scientific implications. We study this problem quantitatively. We use the Wright-Fisher process from population genetic theory to relate the selective coefficient acting on variants to their allele frequency distribution. We use this model to draw conclusions on the allelic frequency spectrum of common diseases and on the power to detect rare variants for different selective coefficients and demographic histories,

In addition, we propose a two-class model for human variants and use it to analyze exome sequencing data. We provide novel estimators for gene-specific selective coefficient, and for the cumulative heritability explained by rare va