Abstract:
Mycobacterium tuberculosis (Mtb) remains a global challenge that has been worsened by the emergence of drug resistant strains of Mtb. We used publicly available Next Generation Sequencing (NGS) and drug susceptibility (DST) data to develop “Resistance sniffer”, an online software program for the rapid prediction of lineage and Mtb drug resistance. Based on the distribution of polymorphisms in the genomes of Mtb, we calculated the power of association between the polymorphisms in different clades of Mtb and resistance to 13 anti-TB drugs. Our data suggests that the development of drug resistance in Mtb is a stepwise process that involves the accumulation of polymorphisms in the Mtb genome. We carefully curated the polymorphisms based on their association powers to create a diagnostic key that captures patterns of these polymorphisms that can be used to predict lineage and drug resistance in Mtb. This diagnosis key was incorporated into the Resistance Sniffer tool, an online software program that we developed for the rapid diagnosis of drug resistance in Mtb. The tool was tested using sequence data from the South Africa Medical Research Council (SA-MRC). Our data suggests that the majority of the strains in SA may have been brought by the arrival of European settlers while the more resistant strains may have been introduced in the region by Asian travellers later on.
We next sought to determine non-random associations between polymorphic sites in genomes of Mtb. Using the attributable risk (Ra) statistical methods, we distinguished between functional associations and associations that may have been due to genetic drift events for different Mtb clades. We then integrated the (Ra) data with drug susceptibility and annotation data to generate networks in Cytoscape 3.71. These networks were then used to infer evolutionary trajectories that drive the emergence and fixation of the drug resistant phenotype in different clades of Mtb.
We demonstrate that strains from the Lineage 1.2 are associated with less complex functional associations compared to the strains from other clades such as the Asian and Euro-American clades. Our data also shows that the predisposition of strains from the Asian clades to develop multi-drug resistance may be attributed to a complex network of functional interactions of mutations in genes that are involved in several aspects of Mtb physiology such as cell wall modelling, lipid metabolism, stress response and DNA repair.
