ABSTRACT
Tuberculosis (TB) is the leading cause of death globally from an infectious disease. Understanding the dynamics of TB’s causative agent Mycobacterium tuberculosis (Mtb) in host is vital for antibiotic treatment and vaccine design. Here we use longitudinally collected clinical Mtb isolates from the sputa of 307 subjects to investigate Mtb diversity during the course of active TB disease. We excluded cases suspected of reinfection or contamination to analyze data from 200 subjects, 167 of which met microbiological criteria for delayed culture conversion, treatment failure or relapse. Using technical and biological replicate samples, we defined an allele frequency threshold attributable to in-host evolution. Of the 167 subjects with unsuccessful treatment outcome, 16% developed resistance amplification between sampling; 74% of amplification occurred among isolates that were genotypically resistant at the outset. Low abundance resistance variants in the first isolate predicts the fixation of these variants in the subsequent sample. We identify in-host variation in resistance and metabolic genes as well as in genes known to modulate host innate immunity by interacting with TLR2. We confirm these genes to be under positive selection by assessing phylogenetic convergence across a genetically diverse independent sample of 10,018 isolates.
