GENETIC DIVERSITY, POPULATION STRUCTURE AND INFLUENCE ON LIFE-HISTORY TRAITS OF THE AFRICAN CATFISH, CLARIAS GARIEPINUS (BURCHELL 1822), IN KENYA

Abstract

The African catfish, Clarias gariepinus (Burchell 1822) is an important species in fisheries and aquaculture in Africa. In Kenya, farmers use seeds of unknown genetic characteristics. Sourcing of brood stock for propagation at hatcheries is not controlled, with inter-basin transfer of brood stock being common. This study used 427 base pairs (bp) of mitochondrial D-loop sequence markers to determine genetic variation and population structure in 5 natural (Lakes Victoria (LVG), Kanyaboli (LKG), Turkana (LTA), Baringo (LBA) and Jipe (LJP)) and 5 farmed populations (Sangoro Aquaculture Center (SAN), Sagana Aquaculture Centre (SAG), University of Eldoret Fish Farm (UoE), Kibos Fish Farm (KIB), and Wakhungu Fish Farm (WKU)) of C. gariepinus collected across Kenya. Similarly, 6 microsatellite DNA markers were used to determine genetic variation in 8 populations (LVG, LKG, LTA, LBA, SAN, SAG, UoE and KIB).The 5 natural populations had higher numbers of haplotypes compared to the 5 farmed populations. Haplotype diversity values were generally consistent with haplotype numbers, with populations of higher haplotypes recording higher haplotype diversity. 88.2% of haplotypes in the 10 populations was private, with LJP showing the highest number at 12, while WKU had the least with 1. All except LJP and LTA populations shared haplotypes, and KIB had the highest number of shared haplotypes at 8. The 68 haplotypes identified in the 10 populations clustered into 5 groups: LVG LJP, LTA, LBA and SAG, both in the Maximum likelihood tree, and in the haplotype network. A total of 31 of 45 pair wise comparisons of the population differentiation index (FST) values were significantly different (p≤0.05). Microsatellite analysis showed farmed populations of higher number of alleles than natural populations, but higher observed and expected heterozygosity were recorded in the natural populations. The number of private alleles was generally uniform in the populations, although KIB and LVG had higher values. For microsatellites DNA analysis, a total of 15 out of 28 pair wise comparisons of the population differentiation index (FST) values were significantly different (p≤0.05), with most of the variation attributed to individual samples (96.72%). All populations were in Hardy-Weinberg equilibrium, since none had significant values for exact tests of H-W at all loci. The 8 populations grouped into 4 genetic clusters (LVG, LTA, LBA and SAG) in structure analysis, with all farmed populations grouping into the Lake Victoria population, and 3 grouping into LBA. Mean relative fecundity for the three populations was 81.9±6.0, 50.8±5.6 and 53.0±5.1 eggs/g body weight for lakes Victoria, Baringo and Kanyaboli respectively, with relative fecundity being higher in Lake Victoria than Lakes Baringo and Kanyaboli, which had similar values. Size at first maturity was higher in LBA than LVG and LKG, while a higher size at maturity was recorded in LVG compared to LKG. Therefore fecundity of fish seems to correlate with Hetereozygosity, while size at first maturity seems to be influenced more strongly by environmental factors than genetic characteristics of C. gariepinus. Water quality parameters were similar among the three sites (Lakes Victoria, Baringo and Kanyaboli) for nutrients (Total phosphorus and total nitrogen), while the physico-chemical parameters varied significantly (p≤0.05) among sites and months of sampling. The findings suggest that LVG, LTA, LBA, LJP and SAG are genetically distinct populations, which can potentially be exploited for aquaculture. Natural populations had higher genetic variation than farmed populations, possibly due to inbreeding depression from domestication of farmed species. Farmers may increase seed production by using populations of C. gariepinus of higher genetic diversity.