Abstract
Infections with Brachyspira hyodysenteriae, the etiological agent of swine dysentery, result in major economic losses in the pig industry worldwide. Even though microbial differentiation of various Brachyspira species can be obtained via PCR, no quick diagnostics for antimicrobial susceptibility testing are in place, which is mainly due to the time-consuming (4 to 7 days) anaerobic growth requirements of these organisms. Veterinarians often rely on a clinical diagnosis for initiating antimicrobial treatment. These treatments are not always effective, which may be due to high levels of acquired resistance in B. hyodysenteriae field isolates. By using long-read-only whole-genome sequencing and a custom-trained Bonito base-calling model, 81 complete B. hyodysenteriae genomes with median Q51 scores and 99% completeness were obtained from 86 field strains. This allowed the assessment of the predictive potential of genetic markers in relation to the observed acquired resistance phenotypes obtained via agar dilution susceptibility testing. Multidrug resistance was observed in 77% and 21% of the tested strains based on epidemiological cutoff and clinical breakpoint values, respectively. The predictive power of genetic hallmarks (genes and/or gene mutations) for antimicrobial susceptibility testing was promising. Sensitivity and specificity for tiamulin [tva(A) and 50SL3(N148S), 99% and 67%], valnemulin [tva(A), 97% and 92%), lincomycin (23S(A2153T/G) and lnuC, 94% and 100%), tylvalosin (23S(A2153T/G), 99% and 93%), and doxycycline (16S(G1026C), 93% and 87%) were determined. The predictive power of these genetic hallmarks is promising for use in sequencing-based workflows to speed up swine dysentery diagnostics in veterinary medicine and determine proper antimicrobial use. IMPORTANCE Diagnostics for swine dysentery rely on the identification of Brachyspira species using molecular techniques. Nevertheless, no quick diagnostic tools are available for antimicrobial susceptibility testing due to extended growth requirements (7 to 14 days). To enable practitioners to tailor antimicrobial treatment to specific strains, long-read sequencing-based methods are expected to lead to rapid methods in the future. Nevertheless, their potential implementation should be validated extensively. This mainly implies assessing sequencing accuracy and the predictive power of genetic hallmarks in relation to their observed (multi)resistance phenotypes.