Testing genetic models for multiple symptoms: an application to the genetic analysis of liability to depression.

A model is presented which allows for the contribution of genes and environment to categorical data on multiple symptoms. The model distinguishes between parameters needed to express the relationship between a latent trait and observed responses and the parameters required to represent the causes of variation in the latent trait. The regression of the latent trait on covariates may also be specified. The model is applied to symptoms of depression in 1983 pairs of adult female monozygotic and dizygotic twins. A model which allows only for polygenic variation in the latent trait is supported as well as the "mixed model," which also allows for the effects of a major gene. The likelihood is significantly lower when all genetic effects are ascribed to a single gene. Practical limitations of the method are discussed.