Regressions between relatives

Author:
Publisher website
Google Scholar
Abstract
SUMMARY: A metric character determined by a large number of loci without epistasis is normally distributed. In the absence of linkage the joint distribution in two or more relatives is multivariate normal, so that all regressions are linear and have constant residual variance. In the presence of linkage this is no longer true except in the case of parent and child; for all other types of relatives the regression line is unaffected by linkage but the residual variance about this line is no longer constant but increases away from the mean.

This publication has 1 reference indexed in Scilit: