Using Digital Twin Simulations to Optimize Genetic Selection in an Admixed Beef Herd

Abstract Polled cattle, or those that are naturally hornless, are highly advantageous from both an economic and animal welfare perspective. While there are a number of different mutations that result in a polled phenotype, a duplication on bovine chromosome 1, referred to as the Celtic mutation, is the most common within beef cattle. Tropically adapted beef cattle typically have some degree of indicine influence, which are all naturally horned breeds. Polled genes must therefore be introgressed into tropically adapted cattle from naturally polled breeds, such as Angus, carrier of the Celtic mutation. However, Angus are not as well suited for hot, humid climates and introgression cannot happen at the expense of productivity or adaptation. Therefore, strategies must be developed to balance the introgression of poll alleles into a population with continued genetic improvement for economically important traits. This study utilized a privately owned, commercial population of admixed cattle to address this challenge. These cattle (n = ~3,000; single year cohort) were previously genotyped and SNP effects were estimated for heifer pregnancy as part of the internal genetic evaluation of the private enterprise. Management and selection of male and female replacements were modelled using actual production parameters in a digital twin framework. Four different breeding systems, multi-sire natural service, artificial insemination, and invitro fertilization collecting either five or 10 oocytes, were modelled and compared with account for differences in selection intensity. Selection index weighting of poll vs heifer pregnancy was modelled in eight different combinations. Genetic crosses were then simulated using the real genotypes in the genomicSimulation package in R for 5, 10, and 30 years of crosses to evaluate long and short-term consequences. We found that this approach of combining digital twin modelling with simulated crosses can be used to identify the selection strategy that should be applied to optimally balance selection for poll and economically important traits, such as fertility. This modelling approach provides a framework that may be customized to fit specific production parameters or genetic populations representing individual operations. More generally, this particular study used cattle and management practices representative of the northern Australian beef industry, and may be used as a reference for producers in that production system of strategies for increasing the rate of polled animals in their herds. Furthermore, this approach may be applied to the selection for or against other single gene traits, such as many genetic disorders. The USDA is an equal opportunity provider and employer.