Fig 1.
Schematic illustration of mutational paths and events.
Four mutational paths that each confer resistance to isoniazid are shown as symbols (see legend at bottom) to the right of a hypothetical phylogenetic tree for 21 MTB strains. Full symbols represent derived genotypes, whereas empty symbols represent ancestral genotypes. The full symbols on the tree represent the reconstruction of the mutational history of the sample. The well-known S315T mutation in katG, encoded by a C > G transversion, is found in eight strains and, in this hypothetical reconstruction, has evolved independently five times. Thus, there are five events for this one mutational path. MTB, M. tuberculosis.
Fig 2.
Transition bias in mutational paths and mutational events in the Basel and Manson data sets.
Symbols represent transition:transverion ratios (Basel paths: 74:78, empirical P value = 7.1 × 10−5; Manson paths: 88:120, empirical P value = 4.2 × 10−3; Basel events: 1,755:1,020, empirical P value < 10−6; Manson events: 1,771:900, empirical P value < 10−6). Error bars represent 95% binomial confidence intervals. The dashed horizontal line shows the null expectation of the transition:transversion ratio, assuming our default null model that one transition occurs for every two transversions and that all mutations are independent. For additional null models used at the level of events, see the main text. The data visualized in this and all subsequent figures are presented in numerical form in S1 Data.
Fig 3.
Relative rates of the six nucleotide pair mutations for mutational paths and events in the Basel and Manson data sets.
Transitions are indicated with bold text. Rates adjusted for GC content (Materials and methods).
Table 1.
Summary of transition bias in mutational events per antibiotic in the Basel and Manson data sets.
Rows are ordered by decreasing number of events in the Basel data set. Dashes indicate antibiotics for which there are no mutational events in the respective data set. Mutations that confer resistance to multiple antibiotics are reported separately and were not counted among the events conferring resistance to individual antibiotics. P values indicating deviation from the default null model and three revised null models (1: the default null model, 2: accounting for path-level bias, 3: accounting for jackpot mutations, and 4: accounting for both path-level bias and jackpot mutations; see main text). Significance (P < 0.05) is indicated for each test by the number of the corresponding null model (e.g., 1, 2 indicates statistical significance for null models 1 and 2).
Table 2.
Observed transitions and transversions in mutational events in the Basel and Manson data sets and in mutational paths in the TBDReaMDB data set among amino acid changes that can be caused by both transition and transversion mutations to the same codon.