Point-of-care diagnostics for antimicrobial susceptibility of Neisseria gonorrhoeae are a promising tool to reintroduce previous gonorrhea treatment regimens into the clinic and reduce the increase of antimicrobial resistance in gonocococcal populations. While a single locus target for ciprofloxacin susceptibility performs well, single loci do not predict susceptibility to other antibiotics. For example, reistance to both penicillin and tetracycline can be acquired via plasmid-encoded or chromosomally-encoded mechanisms, and multiple loci contribute to chromosomally-encoded resistance.
In our study, we used conditional genome wide association studies (GWAS) to identify chromosomal loci that best predict penicillin or tetracycline susceptibility after controlling for the presence of plasmid-encoded resistance determinants. For penicillin, we found that a combination of a sequence in penA, which represents the absence of both a mosaic penA allele and the absence of the resistance-associated insertion at codon 345, and the absence of blaTEM predicted susceptibility with high specificity. We also found that wild-type rpsJ codon 57 and the absense of tetM predicted tetracycline susceptibility with high specificity. Prediction of susceptibility using two loci had only moderate sensitivity, and the lab is working on follow up studies to assess the predictive power when additional loci are included.
This project is now published in The Lancet Microbe!