Pan-genome analysis provides much higher strain typing resolution than multi-locus sequence typing

Abstract

The most widely used DNA-based method for bacterial strain typing, multi-locus sequence typing (MLST), lacks sufficient resolution to distinguish among many bacterial strains within a species. Here, we show that strain typing based on the presence or absence of distributed genes is able to resolve all completely sequenced genomes of six bacterial species. This was accomplished by the development of a clustering method, neighbour grouping, which is completely consistent with the lower-resolution MLST method, but provides far greater resolving power. Because the presence/absence of distributed genes can be determined by low-cost microarray analyses, it offers a practical, high-resolution alternative to MLST that could provide valuable diagnostic and prognostic information for pathogenic bacterial species.

Fig. 1.

Distribution of genes among Staph. aureus genomes.

Fig. 2.

The relationships among Staph. aureus genomes estimated by the NG method. The members of a group are enclosed in a box. Solid arrows are drawn from a genome to its nearest neighbour, distances are indicated next to the arrows. The distance units are as described in Methods and elsewhere in the text. Dashed lines connect members of different groups that are valid neighbours, associating those groups into complexes. Only a single dashed line is drawn between any two groups. (a) Groupings estimated from distances calculated from core gene similarities. (b) Groupings estimated from distances calculated from the presence or absence of distributed genes.