Mykrobe

Mykrobe predicts antibiotic resistance in Mycobacterium tuberculosis from whole-genome sequencing (WGS) data to support drug susceptibility testing.

Key Features:

• Enhanced Mutation Catalogue: Incorporates an updated mutation catalogue that improves sensitivity for detecting resistance, particularly for pyrazinamide.
• User-Defined Resistance Catalogues: Supports user-defined resistance catalogues to enable customized analyses of emerging resistance patterns.
• Non-Tuberculous Mycobacterial Species Identification: Includes algorithms to identify non-tuberculous mycobacterial species from sequencing data.
• Statistical Model for Nanopore Sequencing Data: Implements a statistical model tailored for Oxford Nanopore Technologies sequencing data to improve compatibility and accuracy.

Scientific Applications:

• Drug Resistance Prediction: Predicts resistance to key tuberculosis drugs with reported performance metrics: rifampicin (100% sensitivity, 99% specificity), isoniazid (95% sensitivity, 100% specificity), pyrazinamide (82% sensitivity, 99% specificity), and ethambutol (99% sensitivity, 99% specificity).
• Personalized Therapeutic Regimen Design: Guides personalized treatment regimens with reported 94% concordance with phenotypic DST-driven regimens and higher concordance than other benchmarked tools.

Methodology:

Leverages mutation catalogues from the CRyPTIC consortium (2018) and Walker et al. (2015), validated on extensive M. tuberculosis Illumina sequencing datasets, and benchmarked against four other tools across 10,207 samples including initial and independent sets and a prospectively collected dataset for error-rate estimation.