Estimation of genetic model parameters: variables correlated with a quantitative phenotype exhibiting major locus inheritance.

A major locus that is detected through its effect on one phenotype (a primary trait) may also affect other quantitative phenotypes or qualitative disease endpoints (secondary traits). The pattern of effects of the major locus on a set of primary and secondary traits suggests candidate defects for the mutant allele. The effects are directly estimable when "measured genotypes" or a tightly linked marker allow unambiguous assignment of major locus genotypes. When genotypes assignments are ambiguous for a major locus detected through its effect on a quantitative primary trait, we propose estimators using genotypic probabilities. Making certain reasonable assumptions, we demonstrate asymptotic unbiasedness of these genotypic probability estimators of the genotypic means and variances for either the quantitative primary or secondary traits, of the covariances between quantitative primary and secondary traits, and of prevalences for the secondary qualitative traits. An important application of genotypic probability estimators is to define an effect of a major locus that cannot be detected upon analysis of the variable; for example, major locus effects may be defined for hypertension or blood pressure as secondary traits, but not detected as primary traits.