A novel multilocus sequence typing scheme for Morganella spp.: population structure and introduction of quantitative metric for standardization of MLST scheme.

Objectives: Morganella are opportunistic pathogens increasingly reported in clinical settings. While their phylogeny has been recently updated, no multilocus sequence typing (MLST) scheme is currently available. Developing a MLST scheme can be challenging due to the absence of a strict definition for housekeeping gene selection.

Methods: A quantitative statistical estimator the multilocus correlation score (MCS) was defined to measure the discriminatory power of MLST schemes. The MCS is calculated as the sum of the variation ratio between selected loci compared to the whole genome. The average MCS for 25 existing MLST schemes on >3500 strains was calculated. The MLST scheme for Morganella was designed and tested on a collection of n = 328 clinical isolates. This MLST scheme was also validated by using conventional PCR and Sanger sequencing.

Results: MCS calculated on n = 25 tested MLST scheme allows the selection of an optimal set of genes for Morganella spp. We developed a standardized MLST scheme that can be used for all M. morganii and M. sibonii. This scheme is freely available via the Pasteur database: https://bigsdb.pasteur.fr/. In our collection of 328 clinical isolates, a total of n = 110 sequence types (STs) were identified, including n = 63 for Morganella morganii subsp. morganii, n = 18 for Morganella morganii subsp. intermedius and n = 29 for Morganella sibonii. Comparison with phylogenetic analysis indicated a good correlation of clones obtained by phylogeny and ST.

Conclusions: This study provides a MLST scheme for Morganella spp. offering a tool for tracking clonal complexes circulating worldwide.