Relationship Estimation from Whole-Genome Sequence Data
PLoS Genet, 10 (1):
e1004144 (January 2014)DOI: 10.1371/journal.pgen.1004144
Abstract
The determination of the relationship between a pair of individuals is a fundamental application of genetics. The most accurate methods for relationship estimation rely on precise, localized estimates of genetic sharing between individuals. Earlier methods have generated these estimates from high-density genetic marker data. We performed relationship estimation using whole-genome sequence data for 1490 known pairwise relationships among 258 individuals in 30 families along with 46 population samples as controls. Our results demonstrate that complexities specific to whole-genome sequencing result in regions of the genome that are prone to false-positive estimates of genetic sharing. We provide a map of these spurious IBD regions and introduce new methods, implemented in the software package ERSA 2.0, to control for spurious IBD. We show that ERSA 2.0 provides a 5\% to 15\% increase in relationship detection power for distant relationships with whole-genome sequence data relative to high-density genetic marker data.