Comparison of methods for evaluating genetic differentiation of populations by microsatellite markers

Comparative studies of eight methods for evaluation of genetic differences by microsatellite (STR) markers (11 loci) DNA were carried out using the example of seven breed samples (N = 84) of dairy cattle. The range of indicators of genetic diversity of samples was as follows: the number of alleles per locus 3.5-6.2, the number of effective alleles per locus – 2.4-4.3, the Shannon index – 0.95-1.56, the observed and expected heterozygosity – 0.56-0.97 and 0.53-0.75. Evaluation summary of genetic differentiation by group A methods (FST, GST, and GST(NEI)) were 13.4, 10.3, and 11.8% (pperm < 0.001); the differences between the estimates were statistically insignificant. Estimates by group B methods (G'ST(HED), G''ST(HED), DEST) were 36.4, 37.5 and 29.2% (pperm < 0.001); differences between the estimates were also statistically insignificant. Estimates obtained by group B methods were statistically significantly higher than those obtained by group A methods by almost 3 times. Methods of groups A, B, and C (GDN and uGDN) were used to calculate paired genetic distances in the samples. Despite significant dif-ferences in estimates, the Mantel test showed a high degree of correspondence of the genetic distance matrices (RM≥0.97; pperm < 0.001), which was manifested in the projections of genetic relations of breed samples on plane 1 and 2 of the principal coordinates. The first two principal coordinates explained 97-99% of STR variation in the genetic distance matrices. It can be assumed that when assessing the actual differentiation of populations by STR-markers, the methods of group B should be used and among them the statistics of DEST as independent of the level of mean within-subpopulation heterozygosity. In the study of spatial ordination of gene pools of populations, in all probability, it is reasonably to use any method.