Global whole-genome, phylodynamic, and machine-learning analysis of <i>Glaesserella parasuis</i> serovars 2, 5, and 12

ABSTRACT Glaesserella parasuis is a respiratory pathogen of swine and the causative agent of Glässer’s disease. Among the 15 serotypes, serotypes 2, 5, and 12 represent globally disseminated, high-risk lineages characterized by increased virulence and antimicrobial resistance (AMR). To systematically investigate the global epidemiology and molecular basis of high-risk serotypes, we conducted a large-scale comparative genomic analysis. We assembled 1,004 G. parasuis genomes, including 102 newly sequenced isolates from diseased swine across 18 Chinese provinces and 902 publicly available genomes from 16 countries. Pan-genome analysis identified msmX as a novel marker for precise serotype 5/12 differentiation. Bayesian phylogeographic reconstruction then traced the dissemination history of these lineages: a highly antimicrobial-resistant lineage of serotype 2 likely originated in Japan and spread to the Americas via China in the late 1940s, whereas highly virulent lineages of serotypes 5 and 12 emerged in China before dispersing globally. Resistome and virulome profiling revealed distinct risk patterns: serotype 2 isolates carried more antimicrobial resistance genes (ARGs), while serotypes 5 and 12 harbored broader repertoires of virulence factors (VFs). Notably, we identified isolates co-harboring extensive suites of both VFs and ARGs, representing a convergent dual high-risk genomic profile. Furthermore, machine learning models identified signature genes significantly associated with AMR and virulence, which are implicated in pathways, such as cell wall synthesis, capsular polysaccharide production, and carbon source utilization. Taken together, these findings elucidate the global dissemination patterns and molecular foundations of high-risk serovars and provide critical evidence to guide targeted surveillance, clinical antimicrobial stewardship, and rational vaccine development. IMPORTANCE Glaesserella parasuis poses a global threat to swine health, with serovars 2, 5, and 12 representing high-risk lineages due to enhanced virulence and antimicrobial resistance. However, their global spread patterns and genetic basis remain poorly resolved. Through large-scale comparative genomics of 1,004 isolates, we resolved the transcontinental dissemination routes of these lineages and identified msmX as a novel marker to distinguish serotypes 5 and 12. We further uncover high-risk clones co-carrying extensive virulence and resistance gene repertoires. This study provides a population genomic framework for monitoring high-risk G. parasuis strains and informs the development of targeted vaccines and stewardship strategies to mitigate their impact.