Linkage Disequilibrium in Sequencing Data: A Blessing or a Curse?

• Activity Number: 571
• Type: Invited
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Alkes L. Price*+
• Companies: Harvard School of Public Health
• Keywords: linkage disequilibrium ; sequencing ; imputation ; heritability

Abstract

Linkage disequilibrium (LD) in sequencing data facilitates the imputation of variants sequenced at low coverage, with or without an external reference panel. We explore the limits of this approach in very low coverage data, using either standard computationally intensive methods or Gaussian approximations that enable orders of magnitude reductions in runtime.

On the other hand, LD hinders the estimation of components of heritability, which is an important prerequisite to investing major resources in sequencing studies of complex traits. Even in sequencing data with all causal variants perfectly assayed, LD can bias estimates of the heritability explained by all variants if causal and non-causal variants have different LD patterns, as may occur due to different allele frequency distributions. We investigate this bias and describe possible solutions.