Whole genome sequencing for drug resistance profile prediction inMycobacterium tuberculosis

• DOI: 10.1101/401703
• Publication Date: 2018-08-28T14:08:52Z
• Authors: Sebastian M. Gygli; Peter M. Keller; Marie Ballif; Nicolas Blöchliger; Rico Hömke; Miriam Reinhard; Chloé Loiseau; Claudia Ritter; Peter Sander; Sonia Borrell; Jimena Collantes Loo; Anchalee Avihings...; Joachim Gnokoro; Marcel Yotebieng; Matthias Egger; Sebastien Gagneux; Erik C. Böttger
• Keywords: Epidemiology; https://purl.org/pe-repo/ocde/ford#3.01.05; Antitubercular Agents; FOS: Health sciences; Gene; Computational biology; Drug Resistance, Multiple, Bacterial; Peru; Tuberculosis, Multidrug-Resistant; Pathology; 2736 Pharmacology (medical); quantitative phenotypic drug susceptibility testing; Genome; 10179 Institute of Medical Microbiology; Life Sciences; Thailand; 3. Good health; 3004 Pharmacology; Phenotype; Infectious Diseases; whole-genome sequencing; Democratic Republic of the Congo; Medicine; Ethambutol; Switzerland; Genotype; https://purl.org/pe-repo/ocde/ford#3.03.08; 610 Medicine & health; Microbial Sensitivity Tests; Diagnosis, Treatment, and Epidemiology of Nontuberculous Mycobacterial Diseases; 360 Social problems & social services; Mechanisms of Resistance; drug resistance level prediction; Biochemistry, Genetics and Molecular Biology; Health Sciences; Genetics; Humans; Tuberculosis; Molecular Biology; Biology; drug resistance; Whole Genome Sequencing; 2725 Infectious Diseases; Mycobacterium tuberculosis; Nucleotide Metabolism and Enzyme Regulation; Drug resistance; Whole genome sequencing; FOS: Biological sciences; Mutation; Mycobacterium tuberculosis complex; 570 Life sciences; biology; Genome, Bacterial

Abstract Whole genome sequencing allows rapid detection of drug-resistant M. tuberculosis isolates. However, high-quality data linking quantitative phenotypic drug susceptibility testing (DST) and genomic have thus far been lacking. We determined resistance profiles 176 genetically diverse clinical isolates from Democratic Republic the Congo, Ivory Coast, Peru, Thailand Switzerland by DST for 11 antituberculous drugs using BD BACTEC MGIT 960 system 7H10 agar dilution to generate a cross-validated readout. compared results with predicted inferred whole sequencing. Both methods identically classified strains into resistant/susceptible in 73-99% cases, depending on drug. Changes minimal inhibitory concentrations were readily explained mutations identified Using sequences we able predict levels where wild type mutant MIC distributions did not overlap. The utility was partially limited due incompletely understood mechanisms influencing expression resistance. overall sensitivity specificity genome-based 86.8% 94.5%, respectively. Despite some limitations, has high predictive power infer without need time-consuming methods. One sentence summary accurately predicts may replace culture-based future.