Fig. 3: DCCs exhibit a smoking-related mutational spectrum.

a, Branch classification for each branch type. The left-hand tree illustrates the categorization of non-DCC branches into: within patient, which occur downstream of a node in the tree where all of the descendent sequences are from the same patient; tip branches and branches leading to within-patient clade ancestors where some of the evolution may have occurred within a patient; and internal branches where all the change is likely to have occurred in the environment. The right-hand tree illustrates the additional categorization of DCC branches into deep branches that precede the first population expansion. Note that in DCCs, all branch categories are likely to be within patients, not the environment. b, Inferred environmental mutational spectrum of M. abscessus as calculated from the internal branches of the non-DCCs. The mutational spectrum consists of six different mutations, shown in different colours, each in 16 different nucleotide contexts²⁵. The bars show the proportion of the total mutations of that type in that context along the non-DCC internal branches. c, Mutational spectra attributable to mutations occurring (top) within patients with CF (on the basis of longitudinal isolates collected from infected individuals), and (bottom) during expansion and dissemination of the DCCs (on the basis of analysis of the internal phylogenetic branches of the DCCs excluding the deep branches, that is, after the population expansion). Results are shown as the difference from the environmental spectrum. Asterisks indicate significantly different proportions relative to the environmental mutational spectrum, calculated through permutation analysis. d, Deconvolution analysis reveals relative contributions of specific mutational processes to each spectrum: unassigned (grey), spontaneous (yellow), alkylating agent–associated (purple), ROS-associated (red) and smoking-related (blue).

Source data