The aims of this study were to investigate the role of the diversity of dengue virus populations in changing patterns of virus transmission and disease. Prior to the commencement of this study, dengue 2 virus (DENV-2) had been associated most frequently with severe disease, so the study commenced with this serotype. Because it was not possible to quantitate diversity in the entire 11 kb of the viral genome, the study focussed on the envelope (E) gene, because the E protein is the major protein on the surface of the virion and thus might be under strong selective pressure from the host immune system and from the requirement to engage specific receptors on host cells. This study was the first direct quantification of the diversity of dengue virus populations in individual hosts. The nucleotide sequences of more than 70 per cent of the E genes in each virus population differed from the consensus nucleotide sequence for the population. In the course of quantitating genetic diversity in DENV-2 virus populations in patients and in mosquitoes, recombinant DENV-2 and both parental virus populations were detected in a single mosquito. This was the first such report. In 2001, just after the commencement of this study, Myanmar had the largest outbreak of dengue on record. Unlike previous outbreaks, 95 per cent of dengue viruses isolated from patients were of a single serotype, DENV-1. Despite the large number of cases of dengue, the proportion of patients with severe dengue was low. In the light of these observations, the direction of this study changed to focus on DENV-1. Phylogenetic analysis of the E genes of DENV-1 collected before and after the 2001 dengue outbreak suggested that some time before 1998, an early lineage of DENV-1 had become extinct and had been replaced by two new lineages. There was no evidence that these changes were due to selection or to recombination within the E protein genes of the old clade of viruses and the newly introduced viruses. A more detailed analysis was undertaken, of the entire genome of 11 human DENV-1 isolates and of 4 from mosquitoes recovered in Yangon between 1971 and 2002, to determine whether the extinction of the pre-1998 lineage of DENV-1 (clade A) and the appearance of the two new lineages (clades B and C) could have been due to selective pressures acting on genes other than E. Evidence of only weak selection was found in the NS5 gene (at amino acids 127,135 and 669) but the resultant amino acid changes did not distinguish all recent viruses from viruses belonging to the extinct clade. The phylogenetic relationships between individual genes from these viruses and between the open reading frames were similar. No evidence was found of recombination that might have given rise to two new clades of virus with enhanced fitness. Collectively, these data suggested that the extinction of clade A viruses and their replacement by the two new clades, between 1998 and 2000 was a stochastic event in an inter-epidemic period when rates of virus transmission were low. This was the first report of such an extinction of a lineage of DENV-1 and its replacement by new lineages. At about the same time as the 2001 outbreak of DENV-1 infection in Myanmar, an outbreak of DENV-1 began in the Pacific. A comparison of the nucleotide sequences of the E genes of viruses from the Pacific with those of viruses from throughout south-east Asia suggested that the outbreak in the Pacific was due to the introduction of multiple genotypes of DENV-1 from Asia and that some of these DENV-1 could have originated in Myanmar. The principal observations from this study are: - (a) Dengue virus populations in individual hosts are extremely heterogenous and may contain a significant proportion of non-infectious genomes. (b) Intra-serotypic recombination between dengue viruses may be far more common than the literature suggests but it may not be detected because of the almost universal use of consensus nucleotide sequences. (c) Significant changes in dengue virus genotypes that occur at single localities may be due to genetic bottlenecks rather than to selection or to recombination. (d) Dengue viruses can be transported more than 10,000 km to cause outbreaks in non-endemic areas. Key words: Dengue viruses, diversity, recombination, selection, genetic bottleneck

<I>Bean pod mottle virus </I>(BPMV, Genus <I>Comovirus</I>, Family: <I>Comoviridae</I>)is an important virus in soybean (<I>Glycine max</I> (L.) Merrill), causing quality and yield loss due to seed coat mottling and seed weight reduction. Although BPMV has been known in Virginia since 1958 and has always been regarded as causing negligible losses, its impact is changing as BPMV incidence has increased in many soybean growing areas of Virginia and the USA in general. From 1997 to 2001, a total of five BPMV isolates (V-W1, V-W2, V-S98-1, V-S98-15 and V-S01-10) were collected in Virginia and characterized. In this study, the effects of these isolates were studied, alone or with Soybean mosaic virus (SMV, Genus Potyvirus, Family Potyviridae) strain SMV G1, and isolates S98-51 and S98-52, on selected soybean cultivars. Individual isolates of BPMV showed variable symptom severity, and resulted in yield loss of between 40.4 to 58.1%, while SMV caused 23.7% in the most severe interactions. Up to 100% yield loss was realized from double inoculations of selected BPMV and SMV isolates, BPMV V-S98-1 + SMV S98-52 and BPMV S98-15 + SMV S98-52 on Hutcheson and Hutcheson Roundup Ready&#174; (BC5) soybeans, respectively. Time of inoculation, a critical factor in the impact of many virus diseases, affected seed coat mottling in four cultivars and seed weight in two cultivars, in tests with four BPMV isolates and three stages of soybean development. All BPMV isolates inoculated to plants at vegetative stage V1-V3 severely increased seed coat mottling and reduced seed weight than those inoculated at V4-V6 and reproductive stage R1-R3. Seedlings grown from non-mottled seeds germinated more uniformly had fewer thin-stemmed seedlings and grew faster than those grown from mottled seeds. Inoculation of various cultivars and breeding lines showed that there was no correlation between the severity of virus-induced foliar symptoms, relative accumulation of SMV, and extent of seed coat mottling. Thus, by avoiding the presence of BPMV at an early growth stage through proper timing of planting to avoid vectors, proper cultural practices like weed control, use of SMV free seeds, and chemical control, it is possible to greatly improve seed quality and reduce yield losses in soybean. / Ph. D.