Application of Genomic Data to Map Genetic Diversity and Enable Genetic Improvement in African Cattle, with Particular Reference to Smallholder Dairy Cattle

Abstract

<p>In developing countries, the livestock sector plays an important role in achieving food security and alleviating poverty for the rapidly growing human population. Cattle populations are an important part of these resources and have significant socio-economic and cultural importance to the livelihood of farmers in Africa. The advances in genotyping technology, such as high-density marker arrays or even whole-genome sequencing, provide unique opportunities for in-depth assessments of genetic structure and genetic diversity of indigenous cattle populations. Understanding population sub-structures plays an important role in several allied fields, including conservation genetics, association studies, and quantitative genetics. Genomic markers also allow breeders to trace relationships between animals accurately and can thus be useful for facilitating genetic improvement and breeding decisions.</p> <p>This thesis focuses on providing important insights into the genetic diversity and structure of African cattle populations and presents methods and tools for genetic improvement in smallholder crossbred dairy systems in Africa. A brief description of the research chapters of the thesis follows:</p> <p><b>Chapter 1</b> presents a general introduction to cattle domestication and a description of African cattle breeds, followed by a review of smallholder dairy farming based on crossbred cattle in Africa. The chapter also discusses linkage disequilibrium and several statistical approaches that are used for genetic diversity estimation, selection of informative SNP panels for breed proportion and parentage assignment, and methods for genotype imputation.</p> <p>In <b>Chapter 2</b>, the genetic diversity and structure of the African indigenous and dairy breed proportions of crossbred cattle populations were studied in relation to<i> Bos indicus</i>, European Bos taurus, and African <i>Bos taurus</i> reference populations using medium-density SNP data. I found that all African indigenous cattle populations are hybrids between<i> Bos indicus</i> and African <i>Bos taurus</i> except some West African pure <i>Bos taurus</i> breeds, with West African and South African populations showing a lower<i> Bos indicus</i> content than East African populations. Estimates of effective population sizes declined for all African cattle breeds from a large population at 2,000 generations ago. The largest European dairy proportions found in Kenyan and Tanzanian crossbreds were Holstein/Friesian and Ayrshire with some influence also from Jersey and Guernsey. In Uganda and Ethiopia, the dairy ancestry was mostly from Holstein/Friesian and in Senegal, the dairy proportion were mostly from Monbeliarde and Holstein.</p> <p><b>Chapter 3</b> focused on the selection of informative SNP markers for estimation of total dairy breed proportion and parentage assignment in African crossbred dairy cattle. For estimation of dairy breed proportion, small SNP panels performed better when the highest proportion of markers was selected to differentiate African <i>Bos taurus</i> from European <i>Bos taurus</i> ancestral populations, compared to markers distinguishing<i> Bos indicus</i> from <i>Bos taurus</i>. In all African crossbred populations, unambiguous parentage assignment was possible with ≥300 SNPs for the majority of the panels when parents were sought among all animals with known genotypes, and ≥200 SNPs when parents were sought only among animals known to be a parent of at least one progeny.</p> <p><b>Chapter 4</b> investigated the inference of local and global ancestry and estimated the heterozygosity proportions in West African crossbreds, and genotype imputation accuracies in African indigenous and crossbred cattle populations. Two approaches were tested to estimate the admixture of crossbred cattle, assuming either two or three ancestral populations. Estimates from both the LAMP-LD and ADMIXTURE approaches were highly correlated (r ≥0.981) and showed an average European <i>Bos taurus</i> content of ~49.7% with SD of ±19.8%. The observed heterozygosity proportions in putative F1 crosses were much higher than for other crossbred and pure breed individuals. The imputation accuracy was generally higher when the reference data came from the same geographical region as the target population, and when crossbred reference data was used to impute crossbred genotypes. The lowest imputation accuracies were observed for indigenous breeds.</p> <p>Finally, <b>Chapter 5</b> summarises the overall conclusion of all the research chapters and a general perspective on the findings and their relevance to the field, and highlighting the main findings of the thesis.</p>