Desirable features for any antibiotic resistance catalogue

In the past few years a growing number of catalogues containing mutations associated with resistance (and susceptibility) to different anti-TB drugs have been published. Some are supplements to papers, some can be found in a version controlled repository and others are a mixture.

The publication by the World Health Organisation of their first Catalogue of mutations in Mycobacterium tuberculosis complex and their association with drug resistance in July 2021 has brought the field much needed credibility, however it has also brought to light some problems.

I expect most people to agree with the statement that any catalogue must be straightforwardly usable by a public health official with limited bioinformatic training and experience, otherwise it will never be adopted in the countries where it stands to bring the biggest benefit.

This naturally leads us to propose that to be unambiguous and easy to understand and deploy a catalogue of mutations should…

• be with respect to a specified version of a reference genome (currently H37Rv version 3 for M. tuberculosis)
• be a single object that is readable by both humans and computers
• contain all the logic necessary to build an antibiogram
• provide an estimate of the uncertainty underlying each association and, ideally, the evidence supporting the association
• be versioned to allow for bugs/mistakes to be corrected following publication
• ideally, be provided in a standard format so that the performance of different catalogues can be directly compared

Unfortunately no published catalogue, including the first WHO catalogue (which I played a small part in collating and creating so must take some of the blame), yet meet the majority, let alone all, of these criteria. These criteria first saw the light of day at a talk I gave at the ESM Annual Meeting in Tirana, Albania in June 2023.

I post the list here so we can revisit it when future catalogues are published and see how they fare.