Drug resistance prediction for Mycobacterium tuberculosis with reference graphs

Abstract

Tuberculosis is a global pandemic disease with a rising burden of antimicrobial resistance. As a result, the World Health Organization (WHO) has a goal of enabling universal access to drug susceptibility testing (DST). Given the slowness of and infrastructure requirements for phenotypic DST, whole-genome sequencing, followed by genotype-based prediction of DST, now provides a route to achieving this. Since a central component of genotypic DST is to detect the presence of any known resistance-causing mutations, a natural approach is to use a reference graph that allows encoding of known variation. We have developed DrPRG (Drug resistance Prediction with Reference Graphs) using the bacterial reference graph method Pandora. First, we outline the construction of a Mycobacterium tuberculosis drug resistance reference graph. The graph is built from a global dataset of isolates with varying drug susceptibility profiles, thus capturing common and rare resistance- and susceptible-associated haplotypes. We benchmark DrPRG against the existing graph-based tool Mykrobe and the haplotype-based approach of TBProfiler using 44 709 and 138 publicly available Illumina and Nanopore samples with associated phenotypes. We find that DrPRG has significantly improved sensitivity and specificity for some drugs compared to these tools, with no significant decreases. It uses significantly less computational memory than both tools, and provides significantly faster runtimes, except when runtime is compared to Mykrobe with Nanopore data. We discover and discuss novel insights into resistance-conferring variation for M. tuberculosis - including deletion of genes katG and pncA - and suggest mutations that may warrant reclassification as associated with resistance.

Keywords: Mycobacterium tuberculosis; benchmark; drug resistance prediction; genome graphs; reference graphs; software.

Fig. 1.

Drug phenotype counts for Illumina (upper) and Nanopore (lower) datasets. Bars are stratified and coloured by whether the phenotype is resistant (R; orange) or susceptible (S; green). Note, the y-axis is log-scaled. PAS para-aminosalicylic acid.

Fig. 2.

Sensitivity (upper panel; y-axis) and specificity (lower panel; y-axis) of resistance predictions for different drugs (x-axis) from Illumina data. Error bars are coloured by prediction tool. The central horizontal line in each error bar is the sensitivity/specificity and the error bars represent the 95 % confidence interval. Note, the sensitivity panel’s y-axis is logit-scaled. This scale is similar to a log scale close to zero and to one (100%), and almost linear around 0.5 (50 %). The red dashed line in each panel represents the minimal standard WHO target product profile (TPP; where available) for next-generation drug susceptibility testing for sensitivity and specificity. INH isoniazid; RIF, rifampicin; EMB, ethambutol; PZA, pyrazinamide; LFX, levofloxacin; MFX, moxifloxacin; OFX, ofloxacin; AMK, amikacin; CAP, capreomycin; KAN, kanamycin; STM, streptomycin; ETO, ethionamide; LZD, linezolid; DLM, delamanid.

Fig. 3.

Sensitivity (upper panel; y-axis) and specificity (lower panel; y-axis) of resistance predictions for different drugs (x-axis) from Nanopore data. Error bars are coloured by prediction tool. The central horizontal line in each error bar is the sensitivity/specificity and the error bars represent the 95 % confidence interval. Note, the sensitivity panel’s y-axis is logit-scaled. This scale is similar to a log scale close to zero and to one (100%), and almost linear around 0.5 (50 %). The red dashed line in each panel represents the minimal standard WHO target product profile (TPP; where available) for next-generation drug susceptibility testing for sensitivity and specificity. INH isoniazid; RIF, rifampicin; EMB, ethambutol; OFX, ofloxacin; AMK, amikacin; CAP, capreomycin; KAN, kanamycin; STM, streptomycin; ETO, ethionamide.

Fig. 4.

Impact of gene deletion on resistance classification. The title of each subplot indicates the gene and drug it effects. Bars are coloured by their classification and stratified by tool. Count (y-axis) indicates the number of gene deletions for that category. The NA bar (white with diagonal lines) indicates the number of samples with that gene deleted but no phenotype information for the respective drug. TP true positive; FN, false negative; TN, true negative; FP, false positive; NA, no phenotype available.

Fig. 5.

Benchmark of the maximum memory usage (left panels) and runtime (right panels) for Illumina (upper row) and Nanopore (lower row) data. Each point and violin is coloured by the tool, with each point representing a single sample. Statistical annotations are the result of a Wilcoxon rank-sum paired data test on each pair of tools. Dashed lines inside the violins represent the quartiles of the distribution. Note, the x-axis is log-scaled.