ARGENT: ARgentinian GEnomics for Tuberculosis

The biggest cause of deaths due to infectious disease (over 1 million/year) and drug-resistant infections (~0.5 million/year) is the bacterium Mycobacterium tuberculosis, which causes tuberculosis (TB). About one third of deaths caused by drug-resistant infections are due to TB. This is possibly the single biggest infectious public health threat the world faces, with 1.6 million reported deaths globally due to TB in 2017 (https://www.who.int/gho/tb/en/). As drug resistance leads to treatment failure, the standard-of-care is to test a bacterial sample from a patient, to determine which drugs are likely to be effective. The gold standard culture-based approach is slow (taking many weeks) and expensive, so many countries use a fast (hours) test based on DNA amplification to detect resistance to rifampicin, the recommended treatment for non-drug-resistant TB. If they see resistance to rifampicin they pragmatically assume the sample is also resistant to isoniazid, and therefore multi-drug resistant (MDR). When this assumption is wrong, the patient is being subjected to 18 months of inappropriate treatment with brutal side-effects.

There is an alternative: the bacteria causing the infection can be isolated from (usually) the sputum of a patient, and the bacterial DNA decoded by a process called whole genome sequencing. Because drug resistance is caused by specific changes to the DNA sequence (mutations) in the bacterium, sequencing provides a comprehensive output describing the resistance profile of the bacteria. There are many roads to drug resistance, and decoding the whole genome allows access to all of these mutations. A secondary benefit to whole-genome sequencing is that closely related bacteria have very similar DNA, and if one person infects another, their bacterial strains are closely related. Thus one can decode the mutation patterns in TB samples from many patients - this provides actionable information that can direct the prevention and management of outbreaks.

The challenge is turning whole-genome sequencing into a production system that runs in a public health lab or hospital, with acceptable error rates and outputs that make sense to doctors, nurses and public health practitioners. The only country in the world to have adopted whole genome sequencing for TB diagnosis, patient management and outbreak surveillance is the UK.

Through the GCRF-funded CABANA-project, EMBL-EBI is collaborating with the Argentinian National Reference Lab, which is responsible for confirming all the TB diagnoses in Argentina and in neighbouring countries. This project, named ARGENT, will achieve two things.

First, it will enable us to perform a feasibility study of whole-genome sequencing for diagnosis and management of TB in Argentina, working closely with the experts dealing with the TB cases. Introduction of such a service into routine use requires very careful validation and accreditation, which is beyond the scope of this project. However, by specifically introducing a software workflow which is heavily tested, and being developed in the UK into an accredited workflow specifically designed to be shared with other countries.

Second, the project will build a global, open, web platform (Mykrobe Atlas) incorporating all (>50000) global TB genomes, allowing users to upload their own samples and immediately compare theirs with all that have gone before. Since several countries have legal blocks preventing sharing of pathogen sequence data, a key sub-study will be to trial a hub-and-spoke model that allows a country to participate in Atlas while retaining their raw sequence data in country.

Through the supranational network with other countries in Latin America, led by the Argentinian lab, we will share our learning with other national labs in the region, and also in those countries with the greatest burden of disease - in Africa and Asia.

The overarching context for this proposal is that TB is a major global public health threat. There is a desire to try to eliminate TB in South America, which currently has moderate TB burden overall, but with localised hotspots of high burden. In order to achieve this goal, there needs to be an acceleration of progress in diagnosing TB and MDR-TB, appropriately treating cases, and using new analytic tools to detect regions where there is an excess of transmission, for interventions. Also, since infectious diseases do not follow borders, we need improved methods for integrated working between adjacent countries, and with the wider world.

In this proposal we work with the Argentinian National Reference Laboratory for TB, who are the WHO supranational reference lab for detection of drug resistance for Brazil, Venezuela, Perú, Guyana and Paraguay. Our goals are
1.) To translate the benefit seen in the UK by the introduction of routine sequencing of (the DNA of the infecting bacteria in) TB cases, into the Argentinian context. We do this by transplanting the new workflow being developed for accreditation in the UK by a team led by Derrick Crook, based on software from Zamin Iqbal's team.
2.) To use this on a daily basis in the public health context in Buenos Aires, working with the Expert Group for TB which reviews TB cases and potential outbreaks. This process involves working with people over months, developing the standard reporting from sequencing to match their requirements, and literally re-designing and improving the user-interface of our online tool, Mykrobe Atlas, to provide rapid and easy tools that help them do their jobs. The output will be that multiple individuals - doctors, nurses, TB control officers, will have seen the tools and been personally involved in their improvement, and (hopefully) buy into their value.
3.) To use the online tool as a means of sharing data between the National Reference Lab and referring labs, in particular when trying to understand if their are outbreaks and movement of MDR-TB.
4.) To work with the NRL to generate pilot data which can inform the Ministry of Health of the value of routine sequencing. The pre-existing history and experience implicit in the software, inherited from the work done in the UK, and the route to accreditation, remove many concerns about the process and how it can be validated. At the end of this project, we will be able to update SOPs, identify any factors that need improving, and design a proper validation and accreditation process that will be acceptable to the ministry.
5.) Beyond this, this provides a translatable example for the region, and tools for directly analysing output online. The training materials we will produce and the workshops and webinars will allow this experience to be shared.
6.) Finally, by building an enormous public TB service at EMBL-EBI containing over 50,000 genomes will provide immediate value to researchers across the world.
7.) The sub-project on Autonomous Sharing provides a first trial of a method which would allow countries who do not want to , or are legally unable to, share sequence data, to nevertheless participate in global surveillance of TB. Inevitably there will be more work needed to be done on this, but it will be the first pilot of such a method, with potentially major impact.