RESUMO
To form a reference population necessary for genomic selection of dairy cattle, it is important to acquire information on the genetic diversity of the original population. Our report is the first among the studies on breeding of farm animals to implement Wright's F-statistics for this purpose. Genotyping of animals was performed using BovineSNP50 chip. In total, we genotyped 499 heifers from 13 breeding farms in the Leningrad oblast. We calculated Weir and Cockerham's F(st) estimate for all pairwise combinations of herds of breeding farms and the values obtained were in the range of 0.016-0.115 with the mean of 0.076 ± 0.002. Theoretical F(st) values for the same pairwise combinations of herds were calculated using the ADMIXTURE program. These values were significantly (p < 0.05) higher than Weir and Cockerham's F(st) estimates and fell in the range of 0.063-0.136 with the mean of 0.100 ± 0.001. We discuss the reasons for this discrepancy between the two sets of F(st) data. The obtained F(st) values were used to identify reliable molecular and genetic differences between the herds. The ADMIXTURE program breaks the pool of 476 heifers into 16 subpopulations, the number of which is close to the number of herds used in the experiment. Results of the comparison between F(st) values obtained using SNP markers with published data obtained on microsatellites are in support of the common opinion that microsatellite analysis results in underestimation of F(st) values. On the whole, the obtained across-herd F(st) values are in the range F(st) data reported for cattle breeds. Results of comparison of F(st) values with the data on the origin of bulls imported from different countries lead to the conclusion on the expediency of the use of F(st) indicators to assess heterogeneity of the herds. Thus, we have demonstrated that use of F(st) data provides the means to assess genetic diversity of cattle herds and is a necessary step in the formation of a reference population for dairy cattle.