Recombination in Bacteria
Despite their asexual mode of reproduction, most bacteria express conserved molecular pathways that enable recombination. Recombination plays multiple key roles in bacterial evolution, and is critical for the emergence of antibiotic resistance and in the control of phase variation. Yet studying recombination in natural populations of bacteria is particularly challenging due to its high rates across the entire genome, which confounds phylogeny and causes fundamental difficulties for classical methods. We have developed a powerful new approach that circumvents the need for phylogenetic reconstruction, and enables rapid quantification of recombination using large scale sequencing data, including both whole-genome and metagenomics sequencing.Inferring bacterial recombination from large-scale sequencing data
We developed the mcorr method which predicts and fits the pattern of correlated synonymous substitutions in a sample of DNA sequences to infer the parameters of recombination. This method opens the door to many novel analyses including the impact of recombination on selection, genome structure, antibiotic resistance, and community-level behaviors, as well as historical patterns of recombination revealed from ancient DNA.
Measuring homologous recombination rates in core and accessory genomes of bacteria
Bacterial genomes exhibit wide variation of gene content within species. Genes that are present in nearly all strains within a species comprise the ‘core’ genome. In contrast, ‘accessory’ genes are present in a fraction of strains and are subject to gene loss or gain via horizontal gene transfer. We extended and applied the mcorr method to quantify and compare homologous recombination rates in core and accessory genes in different species using >100,000 genome sequences. We discovered that in certain species core genes recombine more frequently than accessory genes.
Correlated substitutions are a general signature of homologous recombination
Using simulations and coalescent theory we show how patterns of correlated substitutions develop within populations that exchange DNA by homologous recombination. We consider the impact of selection on the structure of correlations, and derive analytical forms for correlation profiles under different coalescent models, including Kingman and Bolthausen-Sznitman models. We show how fitting the theoretical predictions can be used to determine the amount of homologous recombination that occurred in a given sample, and to infer the coalescent parameters of the genetic pool with which recombination has taken place.
