Evaluation of the impact of mitochondrial variation in the estimation of breeding values for dairy cattle

Abstract: Mitochondria are independent cellular components responsible for cellular respiration. Through oxidative phosphorylation they convert Adenosine diphosphate and inorganic phosphate into Adenosine Triphosphate, ATP, the essential molecule sourced by all intracellular metabolic processes. As a cytoplasmic component, mitochondria are transferred to offspring in a uniparental fashion. The combination of evolutionary events generated a compact, haploid, non-recombining and significantly conserved mitochondrial genome across mammalian species. In cattle for instance, it is composed by around 16 kbp presented in a circular double-strand molecule that encodes for 22 tRNAs, 2 rRNAs and 13 protein-coding genes linked to energy production. It also presents a regulatory non-coding region known as D-loop. For a while, mitochondrial genetic variation was considered under neutral equilibrium. However, an increasing number of studies are connecting mitochondrial polymorphisms to variability in phenotypical expression in many species. Nonetheless, mitochondrial DNA has been recurrently identified as source of phenotypic variation for production traits in dairy cattle. Reports indicate that up to 5% of the phenotypic variation for such traits is regarded to the mitogenome. The real impact of the mentioned findings on breeding practices are yet unknown. Reflecting that 5% is a significant share on phenotypic variation, especially when comparing the length of nuclear and mitochondrial genomes, this project was performed on an attempt to clarify whether mitochondrial effect should be accounted for in the estimation of breeding values for dairy cattle. Considering that the genetic merit is the sum of nuclear and cytoplasmic components, a dairy cattle breeding scheme selecting for one polygenic trait with multiple observations was simulated. Using the R package “AlphaSimR” both nuclear and mitochondrial genomes were obtained from a coalescent simulation and used to simulate breeding activities. Breeding values were estimated under four scenarios: (1) standard repeatability model based on progeny testing; (2) a repeatability model including mitochondrial effect as random effect and based on progeny testing; (3) standard single-step GBLUP (ssGBLUP); and (4) ssGBLUP including mitochondrial effect. Two scenarios were also tested regarding the number of causative loci in the mitochondrial DNA: (1) all segregating sites were causal; (2) only one segregating site was causal. The project highlighted discrepancies between published data and simulated inferences of mitochondrial diversity, indicating that further investigation of the population genetics of the mitochondria is necessary. Results indicate an advantage of accounting for mitochondrial effect on the estimation of breeding values for female dairy cattle, although no impact on genetic gain was observed. Including mitochondrial effect in breeding value estimations may be most beneficial for the selection of females to be used for in-vitro fertilization or embryo-transfer techniques.