Thesis (MScAgric)--University of Stellenbosch, 2001. / ENGLISH ABSTRACT: DNA micro satellite loci express extensive allelic variation making them convenient markers for research in many fields employing population genetic tools, including aquaculture and conservation genetics. Twelve Oreochromis mossambicus populations from wild, captive and introduced sources in Southern Africa were screened for genetic variation at ten CA repeat micro satellite loci. Three of the loci - UNHI04, UNHlll, and UNH123 - were sufficiently well resolved to screen extensively and were interpreted according to a model of Mendelian inheritance. Data was analyzed in terms of genetic structure and levels of genetic variation, the effect of management regime in captivity through successive generations on genetic diversity, and the nature of phylogenetic relationships present between populations. Exact tests, carried out using Monte Carlo type multiple resampling techniques, and F-Statistics were used to detect and quantify genetic structure among the twelve populations. The Exact test X2 (P < 0.001), a FST of 0.27 (P < 0.001), eST of 0.26, RsT of 0.28, and a <l>ST of 0.17 all indicated significant structuring among the populations. The evident genetic structuring endorsed the practice of maintaining the populations as separate genetic stocks, in separate tanks, in order to preserve unique genetic material for aquaculture strain development. Populations also exhibited some significant deviations from Hardy Weinberg equilibrium characterised by an overall reduced heterozygosity across the loci. In microsatellite studies, null alleles are often suggested as major contributors to heterozygote deficits. To test for null alleles, two controlled crosses of 0. mossambicus were made. The progeny from each cross were examined for expected parental allelic ratios at the UNHI04, UNHlll and UNH123 loci. All three loci presented evidence of possible null alleles. Accelerated inbreeding and genetic drift through successive generations in captivity can reduce heterozygosity and gene diversity. To investigate loss of diversity a sample taken from the Bushmans population in 1999 (N = 25) was compared with a Bushmans 2000 sample (N = 36). The comparison highlighted altered allele frequencies, a significant increase in average observed heterozygosity and a non-significant change in average expected heterozygosity using the UNHI04 and UNH123 loci. Calculation of genetic distances and phylogenetic comparisons between the populations provided insight into the degree of management required in conserving genetic diversity in natural populations of Mozambique tilapia. UPGMA and Neighbour-Joining techniques were used to construct phylogenetic trees using Dm and ({)~)2 distance matrices. Clustering of populations appeared to reflect geographic locality of the source populations, however certain populations were not congruent with geography. Mantel tests were used to expose a possible association between genetic distance matrices generated from each individual locus. An association would support a geographic background to population genetic structure. The Mantel tests did not provide conclusive evidence. Mantel tests for association between the combined locus Dm and (81l)2 genetic distance matrices and a geographic distance matrix were similarly non-significant. Multi-dimensional scaling (MDS) plots of Euclidean distance values for Dm and (81l)2 matrices presented a two-dimensional view of the genetic distance data. The degree of similarity with the UPGMA and Neighbour-Joining tree-clustering pattern was higher for the (81l)2 than for the Dm MDS plots. Scatter plots indicated a reliable non-linear correlation between Euclidean distance and genetic distance for the two-dimensional MDS. The micro satellite markers employed in this research provided molecular information needed for complimenting a co-study on quantitative genetic evaluation of the twelve populations. The quantitative co-study provided measures of average length and weight gain indices for the populations based on progeny growth trials. No significant correlation of average heterozygosity (gene diversity) with either average weight or length gain was found. The significant genetic diversity and structure present between the twelve populations provided rationale for implementing strategies to conserve natural 0. mossambicus populations as genetic resources, and manage captive populations for long term maintenance of genetic diversity. / AFRIKAANSE OPSOMMING: Die verstaffing van groot alleliese variasie deur DNA mikrosateliete maak van hulle gerieflike merkers vir navorsing in 'n verskeidenheid van velde wat gebruik maak van populasie genetiese gereedskap, ingesluit akwakultuur en bewarings genetika. Twaalf 0. mossambicus populasies wat verkry was vanuit die natuur, in gevangeneskap en ingevoerdes, van Suidelike Afrika was getoets vir genetiese variasie by tien verskillende CA-herhalende mikrosateliet loci. Drie van die loci - UNHI04, UNHlll en UNH123 - is op grootskaal getoets en volgens In model van Mendeliese oorerwing geinterpreteer. Die data was ontleed volgens genetiese struktuur en vlakke van genetiese variasie, die effek wat bestuur strategie in gevangeneskap op genetiese diversiteit in opeenvolgende generasies uitgeoefen het, so wel as die aard van die filogenetiese verhoudings wat teenwoordig is tussen die populasies. "Exact" toetse is uitgevoer deur gebruik te maak van Monte Carlo tipe veelvuldige hermonsterinsamelings tegnieke en F-statistieke is gebruik vir die deteksie en kwantifisering van die genetiese struktuur tussen die twaalfpopulasies. Die Exact toets X2 (P < 0.001), 'n FST van 0.27 (P < 0.001), SST van 0.26, RsT van 0.28, en 'n <DST van 0.17 gee almal 'n indikasie van betekenisvolle strukturering tussen die populasies. Die genetiese struktuur bevestig die beleid dat die populasies behou moet word as aparte genetiese voorraad, in aparte tenke, om te verseker dat die unieke genetiese materiaal behoue bly om akwakultuur variante te ontwikkel. Populasies het ook betekenisvolle verskuiwings van die Hardy Weinberg ekwilibrium getoon, wat gekarakteriseer word deur 'n algemene verlaging van heterosigositeit oor die loci. Nul allele word dikwels aanbeveel om in mikrosateliet studies groot bydraes te maak tot hetersigotiese defekte. Om vir nul allele te toets was twee gekontroleerde kruisings van 0. mossambicus gemaak. Die nageslag van elke kruising was getoets vir verwagte ouer alleliese verhoudings by die UNHI04, UNHlll en UNH123 loci. Al drie loci het getoon dat dit moontlike nul allele kan wees. Versnelde inteling en genetiese drywing deur opeenvolgende generasies in gevangeneskap kan die heterosigositeit en diversiteit verminder. Om die vermindering van diversiteit te toets was 'n monster van die Busmans 1999 (N=25) populasie vergelyk met 'n monster van die Bushmans 2000 (N=36) populasie. Die vergelyking het veranderde alleel frekwensies, 'n betekenisvolle vermeerding in gemiddelde waargeneemde heterosigositeit en 'n onbetekenisvolle verandering in gemiddelde verwagte heterosigositeit getoon deur gebruik te maak van die UNHI04 en UNH123 loci. Berekening van genetiese afstande en filogenetiese vergelykings tussen die populasies het nuwe insig gegee oor die graad van bestuur wat nodig is om genetiese diversiteit in die natuurlike populasies van 0. mossambicus tilapia te behou.UPGMA en Neighbour-Joining tegnieke was gebruik om filogenetiese bome op te stel deur gebruik te maak van Dm en (OIl)2 afstand matrikse. Populasie bondeling het geblyk om geografiese lokaliteit van die bron populasies te toon, alhoewel van die populasies nie met die geografiese lokaliteit ooreengestem het nie. Mantel toetse is gebruik om 'n moontlike assosiasie tussen genetiese afstand matrikse wat verkry is van elke loci bloot te stel. 'n Assosiasie sou 'n geografiese agtergrond tot populasie genetiese struktuur steun. Oortuigende bewys is nie deur die Mantel toetse verskaf nie. Mantel toetse vir assosiasie tussen die gekombineerde loci Dm en (OJ..l)2 genetiese afstand matrikse en In geografiese afstand matriks was ook onbetekenisvol. 'n Tweedimensionele beskouing van die genetiese afstand data is voorgestel deur multidimensionele skaal (MDS) grafieke van Euclidean afstand waardes van die Dm en (OJ..l)2 matrikse te teken. Die graad van ooreenstemming met die UPGMA en Neighbour-Joining boom samevoeging patroon was hoër vir die (OJ..l)2 as vir die DmMDS grafieke. Verspreiding grafieke het 'n vertroubare nie-liniêre korrelasie tussen Euclidean afstande en genetiese afstande vir die twee-dimensionele MDS grafieke getoon. Die mikrosateliet merkers wat in die studie gebruik was het molekulêre informasie verskaf wat nodig is vir 'n komplimentêre studie oor die kwantitatiewe genetiese evalueering van dié twaalf populasies. Die kwantitatiewe studie het afmetings van gemiddelde lengte en gewig vermeerdering van die populasies verskaf gebaseer op nageslag proewe. Geen betekenisvolle korrelasie van gemiddelde hetersigositeit (geen diversiteit) was getoon met óf gemiddelde gewig óf lengte vermeerdering. Die betekenisvolle genetiese diversiteit en struktuur teenwoordig tussen die twaalf populasies het rede gegee om strategieë te implimenteer om natuurlike 0. mossambicus populasies te konserveer as genetiese bronne en om populasies in gevangeneskap te bestuur vir langtermyn instandhouding van genetiese diversiteit.

Thesis (MScAgric)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: Forming part of a conservation programme, this study was concerned with two species of Cichlid fish (Oreochromis mossambicus and O. ni/oticus), which were brought into contact with each other by unnatural ways. They are now hybridizing to some extent and there is also evidence that the foreign O. ni/oticus may out compete the native O. mossambicus. To cast light on what the current distribution is of both these species and the hybrids in Southern Africa, it is important to identify specimens very accurately. In attempting to find genetic markers to distinguish between two species of Cichlids we tested 20 microsatellite dinucleotide (CAn) repeats during a preliminary study and found five of these promising to exhibit little intra-specific genetic diversity but large genetic variation between species. We amplified these five loci in 145 individuals from 10 populations, which included the two species and their hybrids. Exact sizes of the fragments were determined using an automated DNA sequencer. Between the two species, allele sizes were overlapping, but when data were analyzed by statistical models, the differences could be seen for populations, however on individual level there was overlap between the species. The hybrids were found to be intermediate positioned between the two pure species. Our attempt to assign individuals to populations provided doubtful results. Thus, using this set of markers, populations can be ascribed to one of these species, but not individuals by themselves. / AFRIKAANSE OPSOMMING: As deel van 'n natuurbewarings program, word daar in hierdie studie twee spesies van vis ondersoek was in kontak met mekaar gekom het op onnatuurlike wyse. Hierdie twee visspesies vanuit die CICHLIDAEfamilie (Oreochromis mossambicus en 0. ni/oticus) kan hibridiseer wanneer hul saam voorkom, maar dit is ook bekend dat die uitheemse O. ni/oticus die inheemse O. mossambicus kan bedreig in terme van leefruimte, kos en broeispasie. Om die voorkoms van hibriede tussen die twee spesies te ondersoek in Suider Afrika se varswater opvangsgebiede, is dit baie belangrik om individue baie akkuraat te identifiseer. In hierdie poging om genetiese merkers te vind wat die twee spesies van mekaar onderskei, het ons 20 mikrosateliet di-nulkleotied (CAn) herhalende volgordes op verskillende loci ondersoek. Vyf daarvan het belowend voorgekom om as spesie spesifieke merkers te dien. Die fragmente op die vyf loci is ge-amplifiseer in 145 individue vanuit 10 populasies. Presiese groottes van die fragmente is bepaal met behulp van 'n ge-outomatiseerde DNA volgorde bepaler waarna genotiepes vir elke individu toegeken is. Tussen die twee spesies het alleel groottes oorvleuel, maar wanneer data geanaliseer word met behulp van statistiese metodes, was verskille tussen die spesies duidelik op populasie vlak. Die hibriede het intemediêr tussen die twee spesies voorgekom. Dus met behulp van hierdie stel merkers kan onderskei word tussen die twee spesies op populasie vlak, hoewel individue nie op sig self identifiseer kan word nie.