Genetic and Epigenetic Regulation in Angus and Brahman Cattle

Abstract

Angus and Brahman cattle represent two economically important subspecies of cattle with contrasting phenotypes. The Angus cattle breed is representative of the taurine subspecies and has been bred for excellent meat production traits, and the Brahman cattle breed is representative of the indicine subspecies and has been bred for its ability to thrive in harsh conditions. Knowledge of genetic regulation is fundamental to our understanding of what causes these contrasting phenotypes in cattle breeds. Gene regulatory differences can arise because of genetic and epigenetic differences among the breeds, which can shed light on what contributes to the different phenotypes. Despite this knowledge being crucial to understanding how complex traits are controlled, relatively little is known about genetic and epigenetic regulatory differences between the cattle subspecies. This thesis investigated genetic and epigenetic differences between cattle subspecies by using Angus to represent taurine cattle and Brahman to represent indicine cattle in an effort to elucidate factors responsible for their distinct phenotypes. Enhancers are a key genetic regulatory element, but relatively little is known about this DNA element in the cattle genome. The performance of nine machine learning (ML) models and four DNA representations was evaluated to determine the best combination to predict enhancers across species to cattle using models trained on high-quality enhancers in human and mouse. To evaluate the usefulness of cross-species prediction in general, the ML models were also applied to find pig and dog enhancers. For identifying enhancers in cattle, pig and dog, the combination of convolutional neural networks and one-hot encoding to represent the DNA sequence performed the best. They predicted a similar proportion of enhancers in these genomes as what has been estimated to be the proportion of enhancers in the human genome. Whole genome bisulfite sequencing (WGBS) data was generated from Brahman, Angus and reciprocally crossed progeny using fetal liver samples to investigate differential methylation between the Brahman and Angus breeds. The reciprocal crosses were used to investigate the parent-of-origin effects on DNA methylation to determine what role dam and sire genetics had on the progeny's methylome. As breed-specific reference genomes are available for Brahman and Angus, the impact of reference genome choice was investigated to determine how this affects downstream analyses. The methylation analysis identified tens of thousands of differentially methylated regions (DMRs) that were breed-specific and parent-of-origin-specific. One of the DMRs may be controlling the expression of Dgat1 in a breed and sire-of-origin manner. Genome comparison revealed around 75% of CpGs were shared between Brahman and Angus, with around 5% (~one million CpGs) being breed-specific. Moreover, single nucleotide polymorphisms (SNPs) and structural variants (SVs) between Brahman and Angus were 8-fold (p-value < 0.05) and 1.13-fold (p-value < 0.05) higher in CpGs, respectively, and a quantification bias of 2% was observed when the incorrect reference genome was used for analysis. MicroRNA (miRNA) expression data was generated from the same samples that were used to generate WGBS data. This expression data was used to identify differentially expressed, breed-specific and parent-of-origin-specific miRNAs. Fourteen differentially expressed miRNAs (DEMs) were observed between the breeds, with the dam-of-origin and sire-of-origin comparisons identifying one and five DEMs, respectively. Genes that were predicted to be targets of the DEMs were significantly (p-value <0.05) more likely to be differentially expressed than genes not predicted to be targets of the DEMs. The expression of these miRNAs was then correlated with mRNA expression from the same samples and used to identify gene regulatory pathways that may be under microRNA control. MiRNAs that may be involved in regulating heat tolerance in Brahman and fat gain in Angus were identified, as well as a series of signalling pathways that, through differential gene regulation, may contribute to phenotypic differences between Brahman and Angus cattle. Overall, this thesis identified genetic and epigenetic regions of interest between Brahman and Angus that can help shed light on the causes of the contrasting phenotypes observed between Brahman and Angus cattle. This information will benefit future functional studies that look to pinpoint causative elements controlling these traits.