Figure 2

Comparing greedy-BTC model strains to researchers' selections for the Enterococcus data. Neighbor-joining [10] tree based on MLST data for 139 E. faecium strains. Each MLST scheme has seven genes, these genes were concatenated together and a distance matrix (used by both neighbor-joining and the greedy-BTC method) was constructed based on uncorrected distances between pairs of sequences (i.e. the proportion of sites that differ). The blue triangles indicate the five model strains selected using the greedy-BTC method. Coloured circles indicate model strains selected by six researchers. The inset histogram compares the BTC-scores of the greedy and human selection to 1000 random selections of five model strains.