Genetic variation of the Lake Victoria Nile Perch
Abstract
Recent decline in Nile perch abundance in Lake Victoria as deduced from decreasing catches is a
concern to the fishing industry and conservation biology, as population on the decline is expected to
lose most of its genetic variation and because diversity and survivorship correlate positively, this may
threaten its survival and evolutionary potential. To determine the level to which this decline threatens
the persistence of this species, a study was designed to quantify the level of genetic variation and
population structure (existence of mating subgroups, diagnosed by allele frequency differences) in
Lake Victoria. Further, the number of species of Nile perch in Lake Victoria was investigated, as
previous morphometric and allozyme studies lacked adequate power to clearly identify species. To
identify species differences, variation at mitochondrial Cytochrome C oxidase 1 (CO1), a gene used
as a basis for a global identification system for animals (barcode), was sequenced in a panel of 163
samples comprising individuals from five locations transecting the shoreline of the Kenyan gulf of
Lake Victoria: Sio Port, Dunga beach, Mbita point, Mfangano Island, and Karungu bay, and 21
individuals from Lakes Albert (Uganda) and Turkana (Kenya). Sequences were compared among
samples in Victoria and reference samples from Albert and Turkana, and data from a public sequence
repository (NCBI) of other Lates species in the Asian seas. Results showed the species to harbor no
sequence variation at all within Lake Victoria at the CO1 gene. All samples from Victoria were
characterized by a single invariable haplotype, the same haplotype fixed for Lake Turkana and
matched the only available database sequences of Lates niloticus from Tanzania. Lake Albert,
however, exhibited two haplotypes varying in a single T-C transition, one unique while the other
found in Victoria and Turkana. These results suggested that Nile perch of Lake Victoria comprised
one species, Lates niloticus.
Genetic diversity was quantified using heterozygosity and mean number of alleles, while genetic
structure was analyzed using both differentiation at allele frequencies and multilocus genotypes from
nuclear microsatellites to delineate genetic populations using an individual-based approach that
apportions individuals to their source populations using genotypes and Bayesian modeling. Genetic
variation across the seven microsatellites was low with average number of alleles per locus = 5.
Observed heterozygosity (HO) for Victoria populations was 0.848, while the rarefacted mean number
of alleles (allelic richness) ranged from 2.6 to 3.3. HO was consistently and significantly higher than
HE for all populations and overall (p < 0.05). Patterns of allelic distribution showing fewer alleles
relative to a wider allele size range suggested a recent population bottleneck, apparently from recent
introduction of few founders, perhaps confounded by a limited sampling of a narrow geographic
range for this study. Genetic differentiation was low (FST = 0.054; p ≥ 0.05) indicating an absence of
genetic structure. Similarly, Bayesian analysis of genetic structure showed all samples to belong to
one population. In order to ascertain whether variation resided within quantitative traits previously
shown to have adaptive significance in other aquatic species, population divergence and
differentiation was analyzed at meristic and morphometric characters. The number of gill rakers,
counts of rays in pectoral fins and a suite of morphometric characters showed some level of
divergence between two populations delineated from a pool of samples from Lake Victoria and two
reference lakes, Albert and Turkana. Despite appearing to diverge at certain morphological traits, the
evidence of low genetic diversity and a lack of detectible genetic differentiation among populations
is a theoretical pointer to potential loss of alleles, usually affecting the long term persistence.
However, rangewide sampling will be required to confirm these conclusions. This work was
motivated by the need to quantify the extent to which reduction in census has impacted on its genetic
variability. In view of current findings, measures aimed at protecting the perch should consider the
heterozgygote excess as a healthy outcome, but the low sub-population differentiation suggests to a
single species within the lake, and that all geographic regions can be considered under the same
management regime.
Citation
Doctor of Philosophy in Animal Genetics and BreedingPublisher
University of Nairobi Department of Animal Production