Incorporating sequence variants to improve genomic prediction for reproductive traits in Nellore cattle

• Grant number: 20/03975-6
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: January 01, 2021
• End date: December 31, 2022
• Field of knowledge: Agronomical Sciences - Animal Husbandry - Genetics and Improvement of Domestic Animals
• Principal Investigator: Lucia Galvão de Albuquerque
• Grantee: Gabriel Soares Campos
• Host Institution: Faculdade de Ciências Agrárias e Veterinárias (FCAV). Universidade Estadual Paulista (UNESP). Campus de Jaboticabal. Jaboticabal , SP, Brazil
• Associated research grant: 17/10630-2 - Genetic aspects of meat production quality, efficiency and sustainability in Nelore breed animals, AP.TEM

Abstract

Causal variants selected from sequence data analysis are expected to increase accuracy of genomic selection. The main objective of this study will be to evaluate the feasibility of using whole genome sequence (WGS) data for genomic prediction for reproductive traits in Nellore cattle using different methods. Additionally, we will perform genome-wide association study (GWAS) using different approaches to select variants for reproductive traits. Reproductive traits to be considered are: scrotal circumference measured in males, age at first calving, heifer pregnancy at first and subsequent breeding service, days to calving and gestation length measured in females. Phenotypic records measured on approximately 250,000 Nellore animals from commercial breeding programs will be used. A set of approximately 2,741 females, 1,500 sires genotyped with the Illumina Bovine HD assay and 1,750 males with 70K SNP markers as well as whole-genome sequence from 150 important sires will be available for the study. The imputation will be performed from 70K to HD and from HD to WGS using FImpute software. After that, the imputed panel will be applied in a GWAS to preselect variants for further use in genomic prediction. For genomic prediction, we will test two scenarios: 1) using HD SNP panel and 2) applying HD panel plus sequence variants preselected from GWAS. Furthermore, we will compare accuracy of prediction using ssGBLUP with unweighted or weighted genomic relationship matrix (G), with linear and non-linear weights. The Bayesian models BayesB and BayesR will also be tested. To evaluate the quality of genomic prediction from different models validation will be carried out for young genotyped animals, with no phenotypes in the reduced data but at least one record in the complete data, using two different approaches: 1) predictive ability as the correlation between phenotypes adjusted for fixed effects and (G)EBV; 2) a method based on linear regressions that is called LR, which uses correlations between (G)EBV in the complete and reduced data as a measure of consistency between subsequent evaluations. The gain in accuracy with genomic prediction methods will be also compared to traditional pedigree BLUP.