EXCOFFIE@cgeuge52.bitnet (LAURENT EXCOFFIER) (01/11/89)
In response to biosci52 about testing heterogeneity of mtDNA allele frequencies. Before trying to test for heterogeneity, did you perform some neutrality tests o n your mtDNA samples ? A recent paper (Mac Rae et al, (1988) Genetics 120:485-494) has shown some evidence for non-neutrality of mtDNA type frequencies in Drosophila pseudoobscu- ra. As for myself, I have carried some homozygosity tests (Watterson, 1978, Genetics 88:405-417) on human mtDNA samples (results not yet published). This test may detect departure from neutrality and/or population equilibrium for a given sample at a given locus. The method consists to compare a statistic F (equivalent to the sample homozygosity for a diploid population, but it is only here to summarize the frequency distribution) calculated on the sample type frequencies and their expectation under a neutral model. The theoretical values of F and confidence intervals may be found in Ewens (1979, Mathematical Population Genetics, Appendix C and D) or obtained empirically after simulation of a series of random samples. This latter method provides you with theoretical mtDNA type frequencies only conditionned by sample size and the number of types in sample. The algorithm for generating random samples of genes is given in Stewart (Appendix of Fuerst et al, 1977, Genetics 86:455-483). Other evidences exist for some mechanism which would reduce mtDNA diversity in most species. The problem is that selection may not be the only factor involved, and I suspect that population size change may account for heterogeneity, as selection should be uniform over a particular species. What do you think ? Laurent Excoffier Laboratoire de Genetique et Biometrie Dept. of Anthropology University of Geneva 1227 Carouge Switzerland