Advancing <i>Legionella pneumophila</i> genomic surveillance with a high-resolution cg/wgMLST schema for outbreak detection and investigation

Abstract Introduction Sequence-based typing (SBT) has been the standard molecular typing method for understanding Legionella pneumophila genetic relationships. However, genome-scale typing approaches, namely core-genome (cg) or whole-genome (wg) multilocus sequence typing (MLST), provide higher discriminatory power. To advance these capabilities, the Legionella International Typing (LIT) workgroup was established to develop, evaluate, and disseminate a novel cgMLST schema with enhanced wgMLST resolution for L. pneumophila investigations. Methods We created and populated the LIT cg/wgMLST schema with chewBBACA software using more than 9000 genome assemblies representative of the species diversity. We applied a multi-step refinement workflow, considering loci prevalence, diversity and presence/absence profile across the species tree, to select the final cg/wgMLST loci, and compared the performance of the LIT cgMLST schema with the previously used 1521-loci schema and assessed its congruence with SBT. Results The LIT schema includes 2009 loci present in 98% of the dataset, forming the static cgMLST schema for routine genomic surveillance, plus 2698 accessory loci for an in-depth wgMLST analysis of clusters of interest. The LIT cgMLST schema maintains moderate agreement with SBT and presents high clustering congruence with the 1521-loci schema, while providing increased resolution. Analysis of epidemiologically related isolates using the LIT cgMLST schema for initial cluster delineation, followed by cluster-specific dynamic wgMLST analysis extending the cgMLST with accessory loci shared among isolates within each cluster, demonstrated increased confidence for outbreak investigation and source identification. Conclusions The LIT schema is expected to contribute to harmonizing genomic surveillance of Legionnaires’ disease at both local and global levels. The schema and associated resources for local implementation are available on Zenodo ( https://doi.org/10.5281/zenodo.17871973 ).