A putative ecocline in Klebsiella pneumoniae

The genetic structure of bacterial species is most often interpreted in terms of demographic processes such as clonal descent but can also reflect natural selection and hence give functional and ecological insight. Klebsiella pneumoniae (KP) disperses effectively around the world and has high recombination rates, which should result in the species having a well-mixed gene pool. Nevertheless, phylogenies based on diverse KP strains contain a "backbone". This structure reflects a component of variation where the first component in Principal Components Analysis (PCA), PC1, explains 16.8% of the total variation. We propose that the component reflects a "bacterial ecocline" generated by diversifying selection on a quantitative genetic trait. We simulated a model in which a trait is influenced by many genes, and strains with the most extreme trait values have a small advantage, which can recapitulate our KP PCA results and other features of its genetic diversity. As well as providing an explanation for the phylogenetic backbone, our results provide insight into how species such as KP can speciate, via stronger selection on the trait or a reduction in gene flow. Our hypothesis that there is a bacterial ecocline in KP raises two questions, namely what the trait is underlying it and why is the trait under diversifying selection? The genes that are most strongly associated with PC1 provide some hints, with the top locus encoding Kpa fimbriae. Identification of the trait, if it exists, should facilitate insight into selection on quantitative genetic traits in natural bacterial populations, which have largely been unstudied in microbiology, except in the atypical context of antibiotic resistance.