A statistical method for adjusting covariates in linkage analysis with sib pairs.

Abstract

We propose a statistical method that includes the use of longitudinal regression models and estimation procedures for adjusting for covariate effects in applying the Haseman-Elston (HE) method for linkage analysis. Our methodology, which uses the covariate adjusted trait, contains three steps: a) modelling the covariate-adjusted population means of quantitative traits through regression; b) estimating the value of covariate-adjusted quantitative traits; and c) evaluating the linkage between the adjusted trait values and the markers based on alleles shared identically by descent. We applied our adjusted HE method and the standard HE method in S.A.G.E. to the sib-pair subset of the Framingham Heart Study distributed by Genetic Analysis Workshop 13 with systolic blood pressure as the quantitative trait. Both methods gave similar patterns for the LOD scores, and exhibited highest multipoint LOD scores near location 70 cM of chromosome 12. The adjusted HE method has two major advantages over the standard HE method used in S.A.G.E.: a) it has the capability to handle longitudinal data; b) it provides a more natural approach for adjusting the repeatedly measured covariates from each subject.