Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Strategies for Using Commercial Data to Inform Genomic Prediction in Swine Breeding
The objective was to quantify the impact of including purebred (PB) and crossbred (CB) phenotypes and genotypes in genetic evaluations on the response to selection of CB performance. Assuming PB and CB performance with moderate heritabilities (h2=0.4), a 3-breed swine crossbreeding scheme was simulated and selection was practiced for 6 generations where the goal of the selection was to increase CB performance. Phenotypes, genotypes and pedigrees for three purebred breeds, F1 crosses and terminal CB progeny were simulated. Selection was performed in PB lines using estimated breeding values (EBV) for each strategy. Strategies investigated were, i) increasing the proportion of CB with genotypes, phenotypes and sire pedigree relationships, ii) decreasing the proportion of PB phenotypes and genotypes, and iii) altering the genetic correlation between PB and CB performance (rpc). Data inclusion strategies were compared to when all PB and no CB data were included. Several selective genotyping strategies to choose CB animals were investigated. Selective strategies included: 1) Random: random selection, 2) Top: highest phenotype, 3) Bottom: lowest phenotype, 4) Extreme: half highest and half lowest phenotypes, and 5) Middle: average phenotype. The ability to accurately estimate variance components using Random, Top and Extreme selective strategies was also investigated. Each scenario was replicated 10 times. Results showed that including CB data in genetic evaluations improved CB performance regardless of rpc or data inclusion strategy. Minimal change was observed in average CB phenotype when PB phenotypes were included or proportionally removed when CB were genotyped. Removal of PB phenotypes and genotypes when CB were genotyped greatly reduced the response in CB performance. Extreme genotyping produced the greatest rate of genetic gain and the highest prediction accuracy. However, Extreme resulted in an over-estimation of phenotypic variance. Results suggest CB performance can be increased by including a minimal proportion of CB records into genetic evaluations, a greater response to selection can be achieved by Extreme genotyping, yet Random genotyping may be preferred for variance component estimation.
See, Garrett, "Strategies for Using Commercial Data to Inform Genomic Prediction in Swine Breeding" (2020). ETD collection for University of Nebraska - Lincoln. AAI28258039.