Distinguishing high mobility genes from a clustering of gene histories: evidence for selection of streptomycin resistance within antibiotic-producing bacteria in soil

A method was developed to calibrate the distance between phylogenetic histories in the space of trees. This distance provides a method for quanti- fying the strength of phylogenetic signals, defines a new test of incongruency between gene histories, allows construction of a gene interrelationship graph which clusters genes with similar (but possibly incongruent) histories, and pro- vides a framework for testing significant differences in gene mobilities. As a demonstration of the method we compared the phylogenetic histories of a set of house keeping genes and candidates for horizontal gene transfer (strepto- mycin resistance gene strA and chitinase chiF) on a set of streptomycete strains that possess all genes. Phylogenetic incongruency was detected even between a number of the house keepers (HKY model), p < 0.1% and thus noise in de- termining phylogenies is not responsible for topological incongruency; however an analysis of the distance between phylogenetic tree histories of the various genes using the nearest neighbour metric and the (L1) distance on tree space shows that the strA gene phylogeny is a significant distance from the house keepers relative to the house keepers amongst themselves, p < 0.5%. Our anal- ysis demonstrates that strA has a distinct evolutionary history separate from that of the house keeping genes trpB, gyrB, recA and 16S rRNA and evolves significantly faster than these genes p < 0.1%, while the evolutionary histories of these housekeepers cluster together. The chitinase chiF was also found to be significantly different from the house keepers p < 1%. Our methods demon-