Inference about genetic correlations

In polygenic systems genetic correlations and the factors and specific genetic variances from genetic correlation matrices are often interpreted in terms of sets of genes common or specific to variables. While these inferences may indeed be true, a genetic correlation is not always sufficient evidence for the inferences. In some cases two variables with all genes in common can have low genetic correlations, and systems with only a few genes in common can have high genetic correlations. The assumptions about genic effects in polygenic systems and their effects on a genetic correlation are explicated and discussed. It is suggested that a distinction be made betweenbiological pleiotropism andstatistical pleiotropism to promote more accurate communication about the genetic associations among traits.Preparation of this paper was supported in part by NICHD Grants HD-07289, HD-10333, and HD-18426, a grant from the Spencer Foundation, and Grant RR-07013-21 awarded to the University of Colorado by the Biochemical Research Support Grant Program, Division of Research Resources, NIH.