Genomic retrospective evaluation of 20 years of selection in Italian Holstein bulls for feet and legs trait

Under strong directional selection,allelefrequencies rapidly change,allowing the identificationof genomic regions carrying genes and variantsthat control selected traits, as production, functional and morphologicaltraits. Here we searched selection sweeps by birth date regression on EBVs and the analysis of changes in allele frequencies. Genomic retrospective evaluation of recent selection wasperformed in 2918 Italian Holstein bulls born between 1979 and 2011. Genotypedata from SELMOL, PROZOO and INNOVAGEN projects were used.Estimated Breeding Value (EBVs) for 32 traitswereprovided by the Italian Holstein association (ANAFI). Bulls were genotyped with BovineSNP50 v.1 and BovineHD SNPchips. SNPs positions were updated to UMD3.1 using SNPchiMp v.3. Genotypes were imputed using BEAGLE (v.3.3.4) to obtain HD genotypes for all individuals. After quality control, a total of 2918 animals and 613,956 SNPs were included in the working dataset. Birth date regressed on Feet and LegsEBVshowsa strong positive trend in the birth date interval analyzed. To detect genomic regions involved, we first identifiedPLUS- and MINUS-variantanimalsfor the target EBV over the total year range (134 bulls, group OVERALL)and within each birth year (130 bulls, group BY_YEAR). Then,SNP allelic frequencies, within each group,wereobtainedfor PLUS and MINUS variantspools and the absolute allele frequency difference (delta)was calculated. Mean delta valueswere estimated in overlapping sliding windows of 50 SNPs.Only windows with the mean delta above the 75th percentile + 1.5*Interquartile rangewere retained. Only overlapping regions between OVERALL and BY_YEAR group were retained. These regions cover the 0.84% of the total windows analyzed.Among these, two regions seem particularly interesting. The ~686Kb region on BTA10 (from position 62,578 to 63,264 Kb) had the highest mean delta on BY_YEAR. The~417Kb region onBTA20(from position 40,738 to 41,155 Kb)had thehighest mean delta on OVERALL.Bioinformatic analysis is underway to identify candidate genes, QTLsand metabolic pathways under selection for this trait.