Analysis of the covariance structure of digital ridge counts in the offspring of monozygotic twins

Abstract

Improved methods for analysis of covariance structures now permit the rigorous testing of multivariate genetic hypotheses. Using Joreskog's Lisrel IV computer program we have conducted a confirmatory factor analysis of dermal ridge counts on the individual fingers of 509 offspring of 107 monozygotic twin pairs. Prior to the initiation of the model-fitting procedure, the sex-adjusted ridge counts for the offspring of male and female twins were partitioned by a multivariate nested analysis of variance yielding five 10 x 10 variance-covariance matrices containing a total of 275 distinctly observed parameters with which to estimate latent sources of genetic and environmental variation and test hypotheses about the factor structure of those latent causes. To provide an adequate explanation for the observed patterns of covariation, it was necessary to include additive genetic, random environmental, epistatic and maternal effects in the model and a structure for the additive genetic effects which included a general factor and allowed for hand assymmetry and finger symmetry. The results illustrate the value of these methods for the analysis of interrelated metric traits.