Computational protocol: Genome-Enabled Prediction of Breeding Values for Feedlot Average Daily Weight Gain in Nelore Cattle

Similar protocols

Protocol publication

[…] The initial dataset consisted of 7236 weight records from the 804 steers, but only those from the 15th up to 77th d in feedlot were considered to estimate ADG, to disregard the first weight, and also because after this period >30% of the animals had already been slaughtered. A linear regression analysis of live weight over time was performed using the remained 3523 records from 803 steers, using the lm function of the R software (). The slope was used as the ADG during the feedlot period for the purpose of considering only the linear weight gain and avoiding comparison with different feedlot period lengths.Steers were assigned to 39 contemporary groups (CG) containing from 5 to 42 animals, which combined information on mating season (three levels), experimental feedlot (two levels), and slaughter group (32 levels of animals slaughtered in the same week). After that, the phenotype and genotype datasets were merged to ensure that they had the same individuals. The summary of age at feedlot entry, starting weight, ADG, and days in feedlot on the remaining animals are presented in .In total, 780 steers and 34 bulls were genotyped with the Illumina BovineHD BeadChip (Illumina, San Diego, CA). The initial dataset contained 742,906 markers, in which unplaced, mitochondrial, and sex-linked SNPs were first discarded, as well as duplicated markers (e.g., two different names and positions for the same SNP). SNPs were also filtered based on two other panels: GeneSeek Genomic Profiler (GGP) HDi 80K and GGP LDi 20K (Gene Seek Inc., Lincoln, NE). The panels were built specifically for B. taurus indicus breeds. Originally, the GGP HDi 80k/LDi 20k contained 74,085/19,721 markers, of which 69,942/18,464 were available in the primary dataset.Paternity correction and quality control (QC) were performed to improve results. To deal with SNPs presenting significant deviation from the Hardy–Weinberg Proportions (HWP) deviation, we checked plots of HWP vs. percentage of heterozygous, and 17 SNPs with >80% of heterozygous were excluded from the three datasets because they could reflect an error during the genotyping procedure (). QC was performed using the R package SNPtats (). SNPs were kept for further analysis only if they had call rate >98% and minor allele frequency (MAF) >1%. The MAF filter excluded 20.0, 1.9, and 7.3% of the total SNPs from the 770k, HDi, and LDi panels, respectively.After QC, the Beagle v.3.3.2 () software was used for phase inference and imputation of missing genotypes for each SNP panel. Finally, to constitute a fourth SNP panel scenario, Tagger (), which is based on linkage disequilibrium (LD) between markers (r2), was used. This tool estimates the r2 between all SNP pairs and then selects a minimal set (TagSNPs) of markers with a r2 ≥ 0.3 with at least one another marker on the same chromosome. We have chosen this threshold because it is the overall average r2 at the distance of 10–25 kb, obtained in a previous analysis of the same animals (). The final number of SNP was 15,863, 63,945, 82,933, and 534,787 for the LDi, HDi, TagSNP, and 770k panels, respectively. […]

Pipeline specifications

Software tools Tagger, SNPinfo
Application GWAS
Organisms Bos taurus, Bos indicus
Diseases Genetic Diseases, Inborn