No description available.

No description available.

Allelic polymorphism at three unlinked loci coding for blood stage proteins of <i>Plasmodium falciparum</i> was studied serologically, using a panel of 27 monoclonal antibodies. The proteins, MSP1, MSP2 and Exp-1, exhibited 39, 8 and 2 serotypes respectively, among 567 Gambian, Nigerian, and Brazilian clinical isolates. In each of two years, within an urban/periurban study area in the Gambia, the observed number of 3-locus combinations was in accordance with expectations assuming random assortment between the loci, i.e. panmixia in the <i>P.falciparum</i> population. A mean of 2.0 <i>P.falciparum</i> clones was detected in patients from this area. Two clones within an individual were more frequently identical at the MSP1 locus than two clones picked randomly from the local population, indicating that non-identical sibling parasites are sometimes acquired from a single mosquito bite. Parasites isolated from children who sleep in the same room are very frequently identical at all three loci, suggesting that a single mosquito may inoculate more than one human on a given occasion. Intragenic recombination in the MSP1 gene accounts for much of the extensive allelic polymorphism detected serologically. Putative epitopes for several monoclonal antibodies are mapped on the basis of allelic sequence-serology correlations. Strong non-random associations between epitopes at different domains of MSP1 are broadly similar in three countries, which could be a result of differential selection on recombinant alleles. The frequencies of polymorphic epitopes of MSP1, MSP2 and Exp-1 remained stable over seven years in the Gambian study area, rare epitopes remained at a low frequency compared to common alternatives, suggesting that some of these polymorphisms are not maintained by frequency-dependent selection. Little evidence was obtained for association between patients' blood groups and merozoite surface protein polymorphisms, although one statistical association between blood group O and epitope 8F6/49 on MSP2 should be tested further.

The main aim of this thesis is the development of methods for analysing data from selection experiments to make inferences about the genetic control of the selected trait. A series of methods for data analysis are developed and applied to both simulated and experimental datasets under infinitesimal (polygenic) genetic models, discrete locus models and mixed inheritance models (which are a combination of polygenic and discrete locus models). The experimental dataset is from a replicated selection experiment on mice in which an F<SUB>2</SUB> population formed from an inbred cross was divergently selected on body weight for 20 generations. The experimental data are initially analysed assuming the infinitesimal model using a Derivative Free Restricted Maximum Likelihood package (Meyer) to produce estimates of genetic parameters. An extension to the package is then developed to allow the variance components of change continuously over time, in effect regressing the variance components on generation number. This method allows for changes to variance components over and above what would be predicted from the infinitesimal model, thereby detecting deviations from the model. When applied to simulated data the method detects no change in additive genetic variance when a polygenic model with a large number of genes (16384) is simulated, but detects significant decreases in the additive variance, as expected, when a smaller number (32) are used. Analysis of the experimental data indicates that the additive and environment variance components increase over the course of the experiment, significantly so in the Low selected lines. Overall there is an estimated increase in phenotype variance of 56% in the Low lines and 14% in the High lines.

A theoretical model of malaria populations is constructed in an attempt to reflect the important features of the human-mosquito-malaria system under the key assumption that strain-specific immunity occurs, and is controlled by alleles at a single locus (to be referred to as the immuno-allelic locus) in the parasite genome. The model is more realistic with regard to the nature of malaria than previous models of strain-structured malaria populations, and includes such features as a short period of immune memory and competition between parasites growing in hosts. The model is complex, and is thus designed to be examined through computer simulation. Both the epidemiological and genetic effects of strain structure on malaria populations are studied. For example, the effects of the number alleles at the immuno-allelic locus on the proportion of hosts infected are examined. Also, the degree of genetic heterogeneity of malaria parasites inside hosts compared to the total level of genetic heterogeneity of the parasite population is examined. This is done by calculating the value of the statistic G<SUB>ST</SUB> for a neutral locus in the parasite genome (referred to as G<SUB>ST</SUB>(n)). The degree of genetic heterogeneity in a parasite population is of interest as it determines the degree to which sexual recombination will lead to the generation of novel genotypes. The generation of such novel genotypes is important as it may allow evasion of a new vaccine, or the development of resistance to anti-malarial drug. In the simulations examined, G<SUB>ST(n)</SUB> is found to decrease (the degree of genetic heterogeneity increases) with increasing numbers of alleles at the immuno-allelic locus. It also decreases with increasing levels of transmission in the population. This latter result appears to agree qualitatively with the findings of recent population-genetic field studies. The importance of taking into account the epidemiology of a parasite population when examining its genetics is also highlighted. Finally, the importance of the effects of a strain structure on methods of control of malaria populations are discussed.

Despite the relatively small scale of the study area and the high dispersal capabilities of red deer, red deer on the Scottish mainland presented high levels of genetic diversity and significant population structure for both genetic markers, microsatellites (F_ST= 0.019; G_ST’ = 014) and mtDNA control region sequences (Φ_ST = 0.3483). The landscape genetics approach indicated that landscape features play an important role in contemporary gene flow of red deer on the mainland, with sea lochs, roads, mountain slopes and forests located along the Great Glen being responsible for most of the genetic differentiation in the study area. Sex-biased dispersal analyses, conducted using both genetic markers (microsatellites and mtDNA), revealed that male-biased dispersal was weak in the study area with male movements probably being predominant at a local scale (between neighbouring estates). In contrast, rarer long distance dispersal events which are more likely to be linked with colonisation of new areas were suggested to be predominantly female-biased. In terms of management, results from this study suggest that past management practices have not strongly affected the genetic integrity, genetic diversity and population genetic structure of red deer on the Scottish mainland. Phylogenetic analyses revealed that none of the red deer individuals included in this study presented a mtDNA haplotype from foreign deer despite the numerous introductions of foreign species of deer in the Scottish mainland such as wapiti (<i>Cervus canadensis)</i> or sika deer (<i>Cervus nippon)</i>. Furthermore, only few localised individuals were found to have potentially descended from translocation events. Results from this study also support the continuation of current policies for the management of red deer man-made landscape features and to avoid further divisions of the group as the might not represent natural boundaries of populations.

The Soay sheep on St Kilda, the free-living population which is the subject of this project, offers an unprecedented opportunity to dissect the architecture of fitness related traits. In this project, 588 animals belonging to a larger pedigree (882 animals) were genotyped to build a complete genetic map comprising 251 markers. Consequently, a whole genome scan was carried out to detect qualitative and quantitative trait loci. Different morphological and physiological traits were chosen on the basis of their relationship to fitness. The loci affecting the Mendelian polymorphisms of horn type and two aspects of coat colour were mapped with good reliability (<i>Horn type </i>mapped to chromosome 10, <i>Coat colour </i>to chromosome 2, <i>Coat pattern </i>to chromosome 13; LOD scores >3). Several quantitative traits were analyzed and one genomic region showed significant linkage with the variation in jaw length (chromosome 11, LOD = 3.59). In addition, suggestive linkages were detected for hind leg length (chromosome 15, LOD = 2.89), birth weight (chromosome 8, LOD = 2.54), and timing of birth (chromosome 2, LOD = 2.70). Finally, two suggestive linkages were scored for resistance to the gastrointestinal parasite coccidian (chromosome 3 and X, LOD = 2.68 and 2.21 respectively). This project is one of the first attempts to dissect the genetic architecture of complex traits in the wild through linkage mapping. The results reported here will improve our understanding of the evolution and genetics of natural populations by allowing a more realistic modelling of the study traits and by opening the way towards the molecular characterisation of the genomic regions of interest.

R.A. Fisher developed a theory of junctions to describe the loss of heterogeneity during inbreeding. When distinct strands of continuous genetic material are brought together by recombination a junction is formed at their boundary. The junction approach simplifies analysis and simulation of multilocus processes involving populations which can initially be described in terms of a finite number of distinct haplotypes. An example from Fisher's work is a sib-sib inbreeding system which initially has four haplotypes. A secondary contact hybrid zone between two populations which have diverged at many loci can be modelled as the meeting of two distinct haplotypes. Barton used this approach to show that for a tension zone at equilibrium between two infinite demes there is a critical value of selection below which loci act independently, and above which loci act in association with each other. A simulation study is used to show that the approach to this equilibrium is slow, such that within reasonable time limits after secondary contact loci act in association even under weak selection. This result is confirmed using an exact solution by Barton. An alternative approximation can be numerically solved for arbitrary selection functions, suggesting a method of dating secondary contact hybrid zones by the progress they have made toward equilibrium. The possibility of applying such a dating method to a natural hybrid zone is explored by extending the original model to consider finite populations, exchange across a continuum, and non-uniform selective effects across the genome.

I present a thorough investigation of the meaning of fitness variance in the context of conditional sex expression, followed by a comprehensive analysis of a weak maternal effects model of conditional sex expression. This includes a consideration of the effects of different sex expression strategies on the population genetics and conservation ecology of species. I focused on three questions: What is the meaning of fitness variance in the context of conditional sex expression? Under what conditions can conditional sex expression evolve? How can we detect conditional sex expression in natural populations, given new findings on its population genetics and the significance of fitness variance? Findings include the following: 1. Fitness variance is not a useful metric to detect conditional sex expression. 2. Conditional sex expression evolves when selection is greater on one sex than the other; this holds true for a wide range of parameter values. This corroborates existing theory. 3. In the first known model including sexual selection and conditional sex allocation, this combination is shown to produce faster adaptation than sexual selection alone. 4. Conditional sex expression has the potential to increase population longevity, and may additionally spread through clade (or group) selection. 5. Conditional sex expression can be reliably detected using a combination of allele frequency trajectories through time and cross-sectional allele frequency data from polymorphic loci; this combination is necessary to avoid false positives from loci under balancing selection.

1. In a stable hybrid zone between two taxa, natural selection creates a barrier to gene flow which counterbalances the homogenising effects of dispersal and interbreeding. Studies of hybrid zones can contribute to our understanding of the nature of reproductive isolation between two populations, and hence of speciation mechanisms. For example, reproductive isolation may be generated by natural selection acting against alleles in the wrong environment or against alleles in the wrong genetic background. In this thesis, I consider these issues with (i) a study of a hybrid zone between the fire-bellied toads Bombina bombina and B. variegata and (ii) development of analytical and simulation models of geographical variation maintained by dines at multiple loci. 2. In a transect across the Bombina hybrid zone in Croatia, the pattern of change in three phenotypic traits (leg length, belly pattern and egg size) corresponds to the stepped dines observed in previous studies of diagnostic allozyme loci. There is close concordance between the mean values of the traits and between estimates of linkage disequilibrium calculated from associations between alternative phenotypic and genetic variables. Clines in allozyme frequency and spot score are wider in males than in females, suggesting differential dispersal patterns. 3. There is direct evidence for hybrid dysfunction in the field: in samples collected from Bombina breeding sites, embryonic mortality, larval mortality and frequency of larval developmental abnormalities all increase significantly towards the centre of the hybrid zone. However, a cohort analysis of adult toads within a central region shows no evidence of differential mortality with respect to genotype. 4. Bombina bombina and B. variegata typically use different breeding habitats, with B. variegata showing strong avoidance of the semi-permanent ponds favoured by B. bombina. The abundance of aquatic predators is shown to differ between the alternative breeding habitats preferred by either species. Behavioural experiments show that the feeding strategy of B. bombina larvae reduces their risk of predation, relative to that of B. variegata larvae, thus demonstrating an adaptive advantage to the adults habitat preference. As the availability of either habitat type changes across the zone, these adaptations imply that differential adaptation across an environmental gradient is also creating a barrier to gene flow between the taxa. 5. An analytical model of multilocus dines maintained by differential adaptation to alternative environments is developed, considering in particular the effect of selection on neutral markers. The dynamics are qualitatively similar to those of previous models of heterozygote disadvantage. Computer simulations are used to test the analytical predictions, and the restrictions imposed by assumptions of weak selection. These show that while dine shape can accurately estimate parameters such as fitness in hybrid populations, others, such as the number of genes under selection, may be less robust. 6. The effect of a habitat preference in a hybrid zone such as Bombina is explored with a simulation model. Pooling across habitats, overall statistical associations increase with preference strength, illustrating the effect of the habitat preference in maintaining the integrity of the parental genomes. However, the resulting magnitude of linkage disequilibrium and deviations from Hardy-Weinberg proportion within habitats vary nonmonotonically with the strength of the preference. The resulting shape of the zone will be largely dependent on the underlying distribution of habitat availability. 7. The Bombina hybrid zone is therefore maintained by both endogenous and exogenous selection. Although theoretical models show that the dynamics of either regime are similar, the results imply that divergence during allopatry has been driven by adaptation to alternative environments, rather than solely non-adaptive factors. The resulting reproductive isolation has therefore been, at least partially, determined by ecological factors, and environmentally-mediated factors such as a habitat preference will generate further barriers to gene flow between two populations.