Fig. 5

Visualization of a subgraph of the pan-genome ccdBG of 341 M. tuberculosis strains (\(k=19\)), corresponding to the end of the RRDR region of gene rpoB. The first \(k-1\) overlapping characters have been omitted from each node and numerical node identifiers were replaced by characters A, B, etc., for clarity. The original subgraph is shown in Additional file 1: Fig. S6. Parallel edges are collapsed into a single edge, shown with its multiplicity. Edge thickness also reflects multiplicity. The reference H37Rv strain follows the path of the dominating edges (i.e., node path ADEFGHIK). Except for the lateral inflow of 21 strains on the right (due to other mutations in upstream regions), we observe that there are three alternative paths from node A to node K: through node B, C or J. These alternative paths are present due to mutations in codons 450 and 452 (“TCG” and “CTG” in the reference), which are shown in green and blue. Specifically, mutations S450L, S450W (not present in Table 11 as it is only observed in 6 strains), and L452P are underlined (codons “TTG”, “TGG”, and “CCG” in nodes B, C, and J, respectively)