New preprint: predicting rifampicin resistance Philip Fowler, 16th August 202416th August 2024 In this preprint we train a series of machine learning models on protein mutations found in rpoB — this is the gene in the M. tuberculosis RNA polymerase complex where mutations can introduce resistance to rifampicin, an important first-line drug in the treatment of tuberculosis. Unlike pyrazinamide, which we have previously published and binds to pncA, the RNA polymerase is an essential gene and therefore resistance-conferring mutations tend to be subtle and, in this case, mostly close to the rifampicin binding site. We find that all the models achieve similar levels of prediction performance and that the most predictive feature is, perhaps unsurprisingly, the distance from the amino acid being mutated to the centre of mass of rifampicin. All the data and code required to create our Test+Train datasets, train the models and produce nearly all the figures in the preprint is available on GitHub. Share this:TwitterBlueskyEmailLinkedInMastodon Related antimicrobial resistance computing publication research tuberculosis
New publication: detecting minor populations important for predicting fluoroquinolone resistance 5th April 20238th December 2023 When predicting if an infection is resistant or susceptible to a specific antibiotic, it is… Share this:TwitterBlueskyEmailLinkedInMastodon Read More
New grant: Ox4TB 17th March 202517th March 2025 Very pleased to announce that I am a co-investigator on the recently announced Oxford4TB project… Share this:TwitterBlueskyEmailLinkedInMastodon Read More
publication New Publication: Lipids can form anti-registered phases 23rd September 201629th September 2018 When we think of lipids phase separating in a cell membrane we usually think of… Share this:TwitterBlueskyEmailLinkedInMastodon Read More