Core genome multilocus sequence typing schemes for epidemiological investigation of Taylorella equigenitalis and Taylorella asinigenitalis.

Taylorella equigenitalis is the causative agent of contagious equine metritis, an internationally regulated sexually-transmitted infection in horses, which is of great concern as it usually results in temporary infertility. Taylorella asinigenitalis, the second member of the genus, is mainly found in donkeys and is considered non-pathogenic, although a first natural outbreak was reported in 2019 in the United Arab Emirates. Multilocus sequence typing (MLST) is currently used to study the epidemiology of Taylorella spp. but, while highly transposable and reproducible, it only focuses on < 0.5 % of the genome (seven genes). We therefore aimed to develop a robust core genome MLST (cgMLST) based on the analysis of 370 T. equigenitalis and 68 T. asinigenitalis genomes belonging to 46 and 18 sequence types (STs), respectively. Typing results based on 1333 loci (84.0 % of the genome) from T. equigenitalis genomes and 1255 loci (80.3 %) from T. asinigenitalis genomes showed that the discriminatory power of both species-specific cgMLSTs was greater than that of MLST, with 368 and 68 distinct core genome STs (cgSTs), respectively. Clustering was congruent between the cgMLST and MLST methods, with few inconsistencies for T. equigenitalis. Maximum allelic distance between epidemiologically-related strains was used to define cgMLST clustering thresholds, set at ≤ 15 and 20 allelic distances for T. equigenitalis and T. asinigenitalis, respectively. These parameters grouped the cgSTs into 47 and 11 clonal groups (CGs), respectively. Overall, the cgMLST method outperformed conventional MLST in distinguishing clonal strains from epidemiologically-linked strains, supporting the hypothesis that typing based on a few housekeeping genes does not always accurately reflect genomic relatedness between strains, and making cgMLST more suitable for outbreak investigations.