Association of genes in bacterial population genomics

This study aims to gain insight into the distribution of genes within a local bacterial population. It is based on genomic and phenotypic data for seventy-two Rhizobium leguminosarum strains isolated from Trifolium repens and Vicia sativa plants growing in a 1 m2 area of roadside vegetation located between Wentworth College and Walmgate Stray on the University of York campus, UK. These were previously classified into five genospecies based on a phylogeny of 305 universal core genes. Because of the incongruence in the phylogeny of genes from different replicons shown in the previous study, the congruence of each rhizobium gene was first analysed by clanistics. Clanistics were applied to each gene tree of the population and can not only identify conserved genes in symbiovar and genospecies but also genes shared between two symbiovars and within genospecies. Genes of the rhizobium population were investigated along with their occurrence patterns in the population using Pearson’s correlation. A broader view of occurrence of genes in the population was illustrated in the gene co-occurrence network, which reflected the organisation of genes with favoured and disfavoured co-occurrence in the population. The computation demonstrated clusters of genes involved in the nodulation process including both annotated and unannotated genes. Due to the diversity of ability to utilise carbon substrates in the rhizobium population, class association rule was chosen as the method to relate the ability to utilise carbon substrates with the distribution of genes in the population. Results demonstrated that there not only exists a relationship between the ability to utilise a substrate and the distribution of genes in the population, but also cooperation of genes involved in the substrate utilisation. The methods discovered genes involved in the utilisation of γ-hydroxybutyric acid, which were consistent with evidence from experiments and the literature. Hence, it can be concluded that gene transfer and loss can cause variation in the gene content of a population, resulting in recognisable sets of genes present in a particular symbiovar or genospecies, or associated with phenotypes such as substrate utilisation.