|
Abstract:
|
Variance components analysis is a powerful and versatile method for studying many phenomena, but quickly becomes computationally impractical when analyzing large numbers of individuals or conducting repeated model fits with iterative likelihood algorithms like REML. Here we describe Population Linkage, a novel application of a method of moments estimator of variance components and their standard errors known as Haseman-Elston regression, which we use to test the relative contributions of genome-wide and local sharing of DNA in a large population cohort toward the variance of a trait, a technique known as genetic linkage analysis. We achieved additional computational savings by pre-processing the input data and re-using intermediate terms in the model fit across multiple genetic regions. We ran Population Linkage on 4 blood lipid traits in over 70,000 individuals from the HUNT and SardiNIA studies, successfully replicating 25 known signals previously found in larger genome-wide association studies which had analyzed more genetic variants. Our results show the potential for linkage analysis and other large-scale applications of method of moments variance components estimation.
|
