The utility of a multivariate normal model for studying familial patterns in medical and psychiatric data.

Quantitative traits (such as blood pressure, height or anxiety level) are usually more similar in related than in unrelated individuals. A positive correlation between two family members indicates that there are common factors (genetic and/or environmental) which predispose to the trait values in this pair of individuals. A negative correlation can indicate that there is (environmental) competition or some other cause for negative interaction between the pair of relatives. Using a flexible method for the analysis of quantitative data measured over pedigrees, it is possible to estimate the magnitude of the correlations between family members as a function of both their genetic relationship and their cohabitation history. This computer-based method can distinguish the effects of shared genes from some of the effects of shared family environment, and can identify negative interactions between family members which are likely to be of particular interest in studies of behavioural and psychiatric traits. The utility of this method is illustrated with pedigree data on blood lead levels, blood pressure levels and psychological traits.