Sliding Window methods to detection of regions under selection in Nellore cattle

Abstract Text: The directional selection in bovines lead to genetic differences between breeds or herds because the frequency of alleles range in function of trait selected. A frequency of alleles favorable to selected traits increasing and together also increase the frequency of alleles in adjacent loci. Therefore, mutations under selection are characterized as regions where the allelic frequencies are different from expected under chance or absence of selection. The objective in this study was identify regions under selection in Nellore cattle through of calculate of the difference in allelic frequencies between herds under divergent selection. To this end, we used three experimental herds belongs to Estação experimental de Zootecnia de Sertãozinho-SP, Brazil, two herd selected for higher yearling weight (NeS and NeT) and one herd control (NeC) selected since 1968 to weight near the average. We used 67 animals from NeC, 153 from NeS and 421 from NeT genotyped by the Illumina High Density Bovine SNP BeadChip® (777K). After quality control were left 456,816 SNPs. We calculated the frequency of allele “B” for each SNP in each herd and the allelic difference in this frequency  between each pair of herds. Then we adopted the method of comparison between populations described by Hayes et al., (2008) termed sliding window average. This one correspond the average of the absolute values of differences to each set of 10 adjacent SNPs where values that exceeding 0.43 are considered significantly higher than expected in no selection. Although the amount of SNPs used in this study is much higher than that used to determine the empirical significance thresholds, the distribution of differences values here calculated show concordance with its patterns. As expected, difference above 0.43 was no observed in the sliding window average from NeS-NeT, if both herds were selected under same criterion and intensity would not make sense the difference between them exceed the threshold established to predict selection effect on the frequencies. In the NeC-NeS and NeC-NeT sliding window averages there were significant values, 74 and 93 respectively, in chromosomes 3, 4, 6, 7, 12, 16, 18, 20 and 24.  The correlation between differences from NeC-NeS and NeT-NeS was high (r=0.62) and the region with max difference between herds contained the PPARGC1A  gene, (Bta6).  Further we will analyze the role of genes around the regions considered under selection and also match these results with putative signature selections detected with other methods.

Keywords: Divergent Selection, SNPs