Diagnosing antibiotic resistance: future trends?

It is Sunday, I’m in Vienna at the European Congress of Clinical Microbiology and Infectious Diseases (ECCMID) congress. As is traditional at all large conventions, I’m sitting on the floor outside one of the halls because the organisers have put one of the most popular sessions (Gram-negative resistance) in one of the smallest rooms. A good time to record some thoughts.

I see a coming paradigm shift away from species to plasmid and, perhaps, resistance genes. Several talks this morning discussed “plasmid outbreaks” in Gram-negatives, where although you see multiple species, what is actually happening is a plasmid is hopping between species. Or worse, resistance genes embedded in a transposon are hopping between plasmids which are themselves hopping between species. The growth and transfer of such beta-lactamase (bla) genes (and probably also mcr-1) will contribute significantly to the growth of antibiotic resistance in the coming years.

This will drive the adoption of using genetics to follow and understand these outbreaks, since culture-based microbiology is species-based and so struggles to sort out what is happening in this sort of situation. Since there are many different genes involved, PCR-based methods will probably not be much help either. Second-generation sequencing (e.g. Illumina), with its relatively short reads will also struggle since it is difficult to tell where a read originates from, although it can tell you if a gene is present or not. That leaves long-read sequencing which is either expensive (e.g. PacBio) or still in development (ONT MinION). So there isn’t yet a cheap, off-the-shelf genetic approach that allows you to sequence bacterial DNA and determine how many and which plasmids are present.

Much of the current whole-genome sequencing effort is focussed on Tuberculosis with good reason: you don’t have to worry about detecting plasmids since there is little horizontal transfer and there is a huge clinical gain to be made in reducing the time to producing an antibiogram. But ultimately, the skills and knowledge here will have to be transferred to Gram-negatives.

So who is going to do it?

The answer to this is, I think, not obvious since there are commercial issues. Basically, any solution will require a sequencing method that produces long reads. The key questions are: (i) will company A who own a long-read sequencing method license it to company B for a “reasonable” amount, if not (ii) will company A develop such a clinical microbiology workflow itself? I worry this will retard the development of a genetics-based microbiology that can follow a resistance-gene or plasmid-mediated outbreak.

You get a very clear illustration of the importance of commercial considerations walking around the Exhibition Hall here at ECCMID; the vast majority of the stands are for diagnostic solutions and there are very few stands marketing antibiotics, even to the point where companies, whose name you would historically associate with antibiotics, appear to only be marketing diagnostics.