Late Breaking Abstract - Multidimensional genomic mapping reveals the pathophysiology of Mycobacterium abscessus

Mycobacterium abscessus (MAB) is an emerging pathogen, particularly for individuals with cystic fibrosis (CF) and has become the most lethal and frequent multidrug-resistant mycobacterial infection in the developed World. MAB infections are increasingly common in CF cohorts, and are difficult and sometimes impossible to treat. Very little is known about the genetic determinants controlling MAB infection, virulence, and antimicrobial resistance, or how mycobacterial and non-mycobacterial genes (derived from horizontal gene transfer) interact to influence pathophysiology. We therefore sought to define the critical pathways regulating MAB biology through multidimensional genomic mapping. We used whole genome sequence data from 2266 respiratory MAB isolates obtained from individuals with CF to map epistatic interactions (using correlation compressed direct coupling analysis). A subset of 350 isolates were phenotyped across multiple dimensions including: antibiotic resistance and tolerance; carbon source utilisation; macrophage infection and death; and host survival and immune response during in vivo Drosophila infection. Experimental and clinical phenotypic data was examined for clustering (using t-distributed stochastic neighbour embedding), and mapped to genomic sequences (by combining computational structural predictions and linear mixed model GWAS). Our study revealed key functional networks controlling the pathophysiology of MAB. For example, we identified known and new determinants of resistance to amikacin, clarithromycin, clofazimine, cefoxitin and linezolid, and defined virulence in Drosophila infection models as a heritable trait which clusters with clinical outcomes in infected patients.