Nonrandom sampling in genetic epidemiology: an implementation of the Hanis-Chakraborty method for multifactorial analysis

Abstract

The method of Hanis and Chakraborty: [Statistics in Medicine 6:629-646, 1987] is known to yield unbiased estimates of familial correlations from family data ascertained according to just about any method of nonrandom sampling. For multifactorial analysis of such family data, the method is implemented here so as to provide a unique solution, handle variable sibship sizes, and provide likelihood ratio tests of hypotheses. Simulation results are presented to illustrate the utility of the proposed implementation.