Testing for homogeneity of Hardy-Weinberg disequilibrium using data sampled from several populations.

Olson (1993, Annals of Human Genetics 57, 291-295) proposed a large-sample test of Hardy-Weinberg equilibrium when genotype data are sampled from several populations with different allele frequencies. The test assumes that a ratio measure of disequilibrium is constant across the populations. In this paper, we consider the problem of testing the assumption of homogeneity of that ratio and propose both a large-sample test and an exact test. The large-sample test is appropriate if sample sizes in all strata are sufficiently large, but is strongly anticonservative if some strata are small. In the latter case, the exact test is preferred and we approximate the P-value of this test using a Markov chain Monte Carlo approach.