New paper: quantitative measurement of effect of mutations on antibiotics in M. tuberculosis

The CRyPTIC project played a major role in the release by the WHO of their first catalogue of resistance-conferring mutations in M. tuberculosis by collecting and collating many thousands of samples. All the available resistance catalogues make a binary prediction (whether a sample is Resistant or Susceptible). One thing CRyPTIC did a bit differently was measure the minimum inhibitory concentration (MIC) of each of 13 different antibiotics using a broth micro dilution 96-well plate — MICs are semi-quantitative in that they follow a doubling dilution and, due to limited numbers of wells, only span a fixed interval of concentrations.

In this paper, Josh Carter and others on behalf of the consortium has analysed the 15,211 samples that CRyPTIC collected and both underwent whole genome sequencing and went on the project’s UKMYC plates so have MIC data available, thereby estimating the quantitative effect of individual mutations on the MIC. This is the first step towards a genuinely quantitative catalogue — this could bring large clinical benefits since some mutations generally accepted as conferring mutations (e.g. c-15t in fabG1) do not, on their own, increase the MIC as much as say S315T in katG. Obviously following further work, one could see how some mutations, when present on their own, could still be treated with the conventional antibiotic at a higher-dose, assuming that is clinically appropriate. Conversely, this work also shows how some mutations that when present on their own lead to a moderate increase in MIC, can lead to a large increase in MIC when both are present in a single sample.