Predicting rifampicin resistance in using machine learning informed by protein structural and chemical features.

Background: Rifampicin remains a key antibiotic in the treatment of tuberculosis. Despite advances in cataloguing resistance-associated variants (RAVs), novel and rare mutations in the relevant gene, , will be encountered in clinical samples, complicating the task of using genetics to predict whether a sample is resistant or not to rifampicin. We have trained a series of machine learning models with the aim of complementing genetics-based drug susceptibility testing.

An improved catalogue for whole-genome sequencing prediction of bedaquiline resistance in using a reproducible algorithmic approach.

Bedaquiline (BDQ) has only been approved for use for just over a decade and is a key drug for treating multidrug-resistant tuberculosis; however, rising levels of resistance threaten to reduce its effectiveness. Catalogues of mutations associated with resistance to BDQ are key to detecting resistance genetically for either diagnosis or surveillance. At present, building catalogues requires considerable expert knowledge, often requires the use of complex grading rules and is an irreproducible process.

Prediction of pyrazinamide resistance in using structure-based machine-learning approaches.

Background: Pyrazinamide is one of four first-line antibiotics used to treat tuberculosis; however, antibiotic susceptibility testing for pyrazinamide is challenging. Resistance to pyrazinamide is primarily driven by genetic variation in , encoding an enzyme that converts pyrazinamide into its active form.

Methods: We curated a dataset of 664 non-redundant, missense amino acid mutations in PncA with associated high-confidence phenotypes from published studies and then trained three different machine-learning models to predict pyrazinamide resistance.