Genetic diversity and admixture analysis of Ethiopian fat-tailed and awassi sheep using snp markers for designing crossbreeding schemes

The aims of the thesis were to select small cost effective set of SNP markers for estimation of admixture level and to identify optimal levels of Awassi admixture for the on-going crossbreeding program that has been run in farmer’s villages in the highlands of Ethiopia since 1997. Improved Awassi (IA) sheep from Israel has been used for the improvement of the productivity of Ethiopian short-fat tailed Menz (M) and Wollo (W) sheep breeds through crossbreeding in Negassi-Amba and Chiro villages. High density ~50KSNP markers distributed along the ovine genome were analyzed for M (n=34), W (n=18), local Awassi (LA) (n=24) and IA (n=23) sheep breeds to assess genetic diversity, population structure, linkage disequilibrium (LD) and levels of inbreeding. Awassi level was estimated for admixed individuals using 74 top FST ranked ancestry informative markers (AIMs). The Awassi admixture level estimates were associated with lamb and ewe performances, body measurements and morphological characters. Spearman’s rank correlation analysis was employed on the 74 AIMs and subsets of top ranked 65, 55, 45, 35, 25 and 15 AIMs in order to assess the possibility of reducing number of markers. Correlation analysis was also employed to associate Awassi levels estimated by SNPs with estimated by farmers. Proportion of polymorphism was highest for LA (96.2%) followed by the Ethiopian breeds (91.7 to 93.0%). Lowest proportion of polymorphic SNP was found in IA (84.3%). Expected heterozygosity was high (0.36) for LA followed by the local Ethiopian (LE) breeds (0.32) and then the IA (0.30). Highest genetic differentiation appeared between LE breeds and IA (FST~0.13) while the two Ethiopian breeds were closely related (FST=0.004). Surprisingly, LA sheep differentiated from IA at a higher level (FST=0.08) than it differentiated from W (FST=0.052) and from M (FST=0.065). Correlation coefficient (r) between individual Awassi level obtained from pedigree information and estimated from 74 SNP data was 0.98 for a small pedigreed research herd. All subsets of SNPs in this study provided consistent genotyping clusters and reproducible results and the top 45, 55, 65 SNPs yielded very similar estimates of ancestry proportions with the 74 SNPs. Mean and standard deviation of proportion of Awassi level in Chiro sheep flocks was 21.1±14.71 and 27.5±17.13% for ewes and lamb, respectively. Whereas, in Negassi-Amba the proportion of Awassi level was much lower with corresponding values of 11.0±10.53 and 9.0±7.36% for ewes and lambs, respectively. Higher within and between population variations were observed in crossbred populations found in the two villages. Best performing 13% cohort lambs at 8 months age had 5.9 and 10.4 Kg higher than the population mean in Negassi-Amba and Chiro, respectively. In general, lamb eight months weight was increased as the Awassi level increased. Lambing interval delayed as the Awassi level increased, however number of lambs weaned ewe-1 year-1 was not significantly different (P>0.05). In both locations productivity of ewe in production of eight months lamb weight year-1 was increased as the Awassi level increased to a certain level. Ear length and hair on leg score were appeared as primary predictors of Awassi level in Chiro (Wollo x Awassi) crossbred population (R2=0.47, P>0.0001 for lambs and R2=0.48, P>0.0001 for ewes). For Menz x Awassi ewes wither height appears as the main predictor which explained 32% of the total variation when alone and explained 42% of the variation when used with hair on leg score. In absence of pedigree and genome information, and when decision on admixture level is not too sensitive, farmers could be able to estimate the admixture level of individuals in their crossbred population. This study provided insights regarding diversity and structure of local Ethiopian and Awassi breeds. Principal component analysis and model based structure analysis revealed clustering of populations according to their geographic location and breed development. Strong selection pressure on the Awassi sheep resulted in significant differentiation from LA. Low level of genetic variability and highest inbreeding level observed within the improved Awassi sheep breed should be noted and steps toward increasing diversity should take in to consideration. About 45 top ranked AIMs selected based on FST were good enough for accurate estimation of the level of ancestry in the crossbred sheep population. Association analysis of Awassi levels with performances suggests crossbreeding of Wollo sheep with Awassi in Chiro and similar areas up to 50% Awassi level and strengthening the ongoing selective pure breeding programs of Menz sheep in Negasi-Amba village. Considering the genetic variation created in the crossbred population, genetic improvement applying continuous selection within the crossbred population along with management improvement should be considered which progressively leads to the development of a composite population. The use of AIMs would inspire livestock breeding programs in developing countries by availing breed composition and pedigree information which persisted as marked constraint.