Exploring the molecular basis of serotyping and antibiotic resistance differences in Riemerella anatipestifer based on pan-genomics and machine learning.

Abstract

Riemerella anatipestifer (R. anatipestifer) is a major pathogen in poultry worldwide, with multidrug resistance (MDR) and diverse serotypes complicating disease control. This study systematically investigated the molecular basis of serotyping and MDR characteristics in 92&#x202f;R. anatipestifer strains isolated from 13 provinces in China between 2008 and 2023 by integrating pan-genome analysis with machine learning approaches. In this study, serotypes 2 (29.4&#x202f;%), 7 (25.0&#x202f;%), and 1 (21.7&#x202f;%) were predominant. All isolates exhibited MDR and strains of serotypes 2 and 7 had significantly higher MICs for tigecycline, doxycycline, polymyxin B, and florfenicol than serotype 1 (P&#x202f;<&#x202f;0.05). Eleven resistance genes were identified, and some (e.g., floR, tetX) displayed serotype-specific distribution patterns. Phylogenetic analysis indicated close relatedness between serotypes 2 and 7, while serotype 1 strains were genetically diverse. The accessory genome was enriched in mobile elements and O-antigen-related genes, and multiple serotype-specific marker genes (e.g., pgIA, wbpI) were identified, revealing the molecular basis of serotype classification. Additionally, insertion sequences (IS1595) flanking CPS gene clusters implied that horizontal gene transfer and recombination events may play pivotal roles in serotype variation. This study provides the first integrative framework combining pan-genomics and machine learning to elucidate the genomic basis of serotype diversity and antimicrobial resistance in R. anatipestifer, offering new insights into its adaptive evolution and informing precision vaccine and diagnostic development.