A putative ecocline in Klebsiella pneumoniae

Are quantitative traits important for bacteriology? The quantitative trait models used in animal and plant genetics assume that the trait takes continuous values and is influenced by many variants circulating in the population, each of which acts approximately independently. Amongst the many continuous-valued bacterial traits such as cell size or antibiotic minimum inhibitory concentration, we do not know of any for which these assumptions have been validated. We propose that such traits can be identified if they are the targets of selection and this selection structures overall genetic variation of the species. The phylogenetic trees of global strain collections of Klebsiella pneumoniae (KP) have a “backbone”, which has not been explained. We propose that this backbone reflects a genome-wide component of variation, which is well-captured by Principal Components Analysis (PCA) applied to diverse KP genomes. We show that this structure can be recapitulated in simulations of diversifying selection on a quantitative trait and cannot be explained by neutral models. An implication is that variation in KP forms an ecocline – i.e. a cline generated by natural selection rather than geography but the trait underlying the cline, if it exists, remains to be determined.