Comparative Population Genomics and Machine Learning Reveal a Distinct Regional Genomic Fingerprint of Middle Eastern...

Acinetobacter baumannii is a major multidrug-resistant nosocomial pathogen, yet its population structure and evolutionary dynamics in the Middle East remain poorly understood. This study compared Middle Eastern and global isolates to identify regional genomic signatures. We analyzed 199 high-quality A. baumannii genomes (100 Middle Eastern and 99 global) using pangenome reconstruction, core-genome phylogenetics, resistome and virulome profiling, genome-wide association studies, and machine learning. The pangenome comprised 10,971 genes, including 1102 unique to Middle Eastern isolates. Compared with global isolates, Middle Eastern genomes carried fewer antimicrobial resistance genes but a higher abundance of mobility-associated genes. A distinct regional cluster was dominated by sequence type 2 (ST2), representing 66% of Middle Eastern isolates. Genome-wide association analysis identified region-associated markers, particularly ISAba1 variants and the tufB gene. A Random Forest classifier distinguished Middle Eastern from global isolates with 79.9% accuracy. Virulence gene repertoires were similar between groups, and resistance gene burdens remained stable over time. Middle Eastern A. baumannii isolates exhibit a distinct regional genomic signature characterized by accessory-genome conservation, reduced genomic heterogeneity, enrichment of ISAba1-associated markers, and increased mobility-associated gene content. These features differentiate the regional population from globally distributed isolates despite the similar prevalence of the ST2/IC2 lineage in both cohorts.