PSII-6 - Selection and evaluation of sequence variant subsets considering multiple types of biological information in beef cattle

Abstract: Four methods were proposed for adding additional criteria to identify informative sequence variations (HISV) from Bos Taurus imputed low-pass genomic sequence data for genetic prediction. All methods selected variants from an original set of ~ 60 million. First, HISV encoding protein cavity sites, protein 3D structures provided by ALPHA Fold 2 were analyzed using the Fpocket application for cavity forming residues and using TBLASTN mapped them to HISV in Bos Taurus genome UCD 2.0. Second, parsed KGML files representing biological pathways of the Bos Taurus (Bos) and Homosapien Sapiens (Human) were collected from the KEGG pathway database and combined based on the metabolites and the proteins shared between them to form a complete representative biological network of Bos Taurus. Through the network centrality analysis method, proteins involved in the organism's critical interactions were identified and sequence variants reported in those proteins were considered HISV. Third, protein family domains from the PFAM database (https://pfam.xfam.org/) were mapped to the Bos Taurus genome using the TBLASTN application to identify HSIV. Finally, 11 different promoter sequence patterns with varying promoter lengths from 7 bases to 10 bases were searched 3000 base pairs upstream of the exons of Bos Taurus to identify the probable promoter sites and the sequence variants reported at the identified promoter sites and 70 bases around the promoter site were considered as HISV. A crossbred beef cattle population (n=2,212) with birth weight (BWT) and post-weaning gain (PWG) recorded was utilized to evaluate the subsets. Heritability was estimated for each subset of variants. Cross-validation was conducted for each variant subset by partitioning animals into biological types of > 25% Indicus, 0-24% British, 25-49% British, 50-74% British, and > 75% British (Indi25, Brit0, Brit25, Brit50, and Brit75, respectively). Cross-validation statistics tested included the Pearson correlation of EBV from testing and adjusted phenotype, and the ratio of correlation of EBV from testing and adjusted phenotype with the square root of heritability where the heritability estimate used was from the subset being evaluated (Ratio). Heritability estimates for BWT ranged from 0.25-0.28 for Bos, 0.19-0.21 for Human, 0.23-0.23 for PFAM, and 0.29-0.31 for promoter sites. Heritability estimates for PWG ranged from 0.36-0.39 for Bos, 0.30-0.33 for Human, and 0.32-0.33 for PFAM, and 0.36-0.37 for promoter sites. Cross-validation results for BWT for Ratio ranged from 0.15-0.31 for Bos, 0.05-0.31 for Human, 0.22-0.34 for PFAM, and 0.09-0.53 for promoter sites. Results for PWG ranged from 0.13-0.43 for Bos, 0.08-0.31 for Human, 0.11-0.41 for PFAM, and 0.27-0.43 for promoter sites. The USDA is an equal opportunity employer and provider.