Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Afrique du Sud → 2017 → Genomics as part of an integrated study of smallholder sheep farming systems in the Western Cape, South Africa

Stellenbosch University (2017)

Genomics as part of an integrated study of smallholder sheep farming systems in the Western Cape, South Africa

Molotsi, Annelin Henriehetta

Titre : Genomics as part of an integrated study of smallholder sheep farming systems in the Western Cape, South Africa

Auteur : Molotsi, Annelin Henriehetta

Université de soutenance : Stellenbosch University

Grade : Doctor of Philosophy (PhD) 2017

The aim of this study was to evaluate the productivity of smallholder sheep farming systems in the Western Cape, South Africa, using an interdisciplinary approach, combining socio-economic factors and genomics. The objectives were firstly to investigate sheep breeding practices followed in smallholder sheep farming systems and to determine socio-economic factors that influence offtake rate of flocks in the Western Cape, South Africa. Secondly, to compare genetic diversity and population structure of smallholder sheep flocks to Western Cape resource flocks. Thirdly, to identify single nucleotide polymorphisms associated with causal variants for the wet-dry phenotype using the ovine 50K SNP bead chip. The final objective was to determine selection signatures within South African sheep populations. To achieve the first objective interviews were conducted with 72 smallholder farmers in three districts of the Western Cape, namely the West Coast, Karoo and Eden. Phenotypic measurements of body weight, body condition score and reproduction (wet and dry phenotype) were obtained from a sample of ewes from the West Coast district from 2011 until 2014. Range condition scores were estimated for two farms in the Karoo and four farms in West Coast district. SURVEYMEANS and SURVEYFREQ in SAS were used to obtain means and frequencies for qualitative data. The production systems were largely small-scale (43-85%) and communal (10-57%). Sheep were mainly kept for consumption purposes. The Dorper was the prominent breed used in all the farming systems and selection emphasis was different for farming systems. Traits selected for by communal farmers included temperament, control of flies, meat taste, growth rate, fertility and foraging ability ; whereas small-scale farmers selected for conformation, mothering ability, disease tolerance, heat tolerance and temperament. Lucerne was the predominant source of feed in addition to natural grazing of rangeland. The major constraints mentioned were land availability, drought and water availability. Average range condition scores were low for both districts and grazed vs. rested camps, implicating overgrazing. Carrying capacity was higher than the recommended carrying capacity in the Eden district (both communal and small-scale farmers, 0.26 and 18.5 ha/LSU respectively) and in West Coast for small-scale farmers. The flock structure showed variation between farming systems with mature ewes (39-86%) and young replacement ewes (0.41-46%), young rams (0.41-15%), rams (0-5%) and lambs (0-2%). The wet-dry phenotype was recorded as a measurement of reproductive success over the three year period. The chi-square test indicated significant differences between reproductive performances over the different years with a repeatability of 0.399. Information on the socio-economic status of the farmers, their production system and reason for keeping sheep was obtained. Information on the average numbers of lambs sold and offtake rate was obtained as an economic indicator. A general linear model was fitted on average number of lambs sold and offtake with district, sources of income, participation in government programs and flock size as fixed effects. The majority of smallholder farmers were above 40 years of age and only 32% ranked their wealth status as poor. The smallholder sheep farming systems in the Western Cape were male dominated. District significantly influenced average number of lambs sold and offtake rate. The Karoo district sold a higher average number of lambs per year (41 ± 8.8), with West Coast and Eden selling the same average number of lambs per year at 7 ± 2.2 and 7 ± 2.6, respectively. Offtake rate for the respective districts were not significantly different and were 17 ± 17% for Eden, 48 ± 20% for the Karoo and 46 ± 20% for West Coast. To achieve the second objective blood samples were obtained from 295 sheep of which 172 had been identified as smallholder Dorpers, 4 smallholder White Dorpers, 46 purebred Dorpers, 26 purebred South African Mutton Merinos and 47 purebred Namaqua Afrikaners. Genetic diversity was estimated using allelic richness (Ar), observed heterozygosity (Ho), expected heterozygosity (He) and inbreeding coefficient (F). Population structure analysis was performed using fastSTRUCTURE to determine the breed composition of each genotyped individual. The Namaqua Afrikaner had the lowest He of 0.280 ± 0.18 while the He of smallholder Dorper, Dorper and South African Mutton Merino did not differ and were respectively 0.364 ± 0.13, 0.332 ± 0.16 and 0.329 ± 0.17. The inbreeding coefficient was highest for the pure breeds, Namaqua Afrikaner, Dorper and South African Mutton Merino compared to the average inbreeding coefficient for the smallholder Dorper population. There was evidence of introgression with Namaqua Afrikaner, South African Mutton Merino and White Dorpers in the smallholder Dorper population. Similarly, the latter population was more genetically diverse than the purebred Dorper, South African Mutton Merino and Namaqua Afrikaner breeds from the resource flock. To achieve the third objective wet-dry records were obtained from smallholder farmers (n = 176) and the Nortier Research Farm (n = 131) for the 2014 breeding season. A logistic regression model was fitted to adjust the data for the fixed effects of farm, breed, and age of the ewe and mating weight as a covariate. Linkage disequilibrium (LD) and inbreeding coefficient were estimated using PLINK. Association analysis was performed using the genome-wide efficient mixed-model association package (GEMMA) to determine whether any significant SNPs were associated with the wet-dry reproductive trait.


Version intégrale (3,01 Mb)

Page publiée le 20 décembre 2018