Global phylogenomic structure and accessory resistance dynamics of Acinetobacter baumannii ST2

Acinetobacter baumannii sequence type 2 (Pasteur ST2), associated with international clone 2, is a major antimicrobial-resistant lineage, but its within-lineage population structure and accessory resistance variation remain incompletely resolved. We analyzed 5612 quality-filtered public genomes from an existing global dataset using PopPUNK, core-genome SNP analysis, AMRFinderPlus, MobileElementFinder and MOB-suite. A dominant PopPUNK cluster contained 5604 genomes (99.86%). Its SNP phylogeny showed dense short-distance relationships within a connected lineage; pairwise distances had a median of 314 SNPs. Four antibiotic classes-beta-lactams, aminoglycosides, phenicols and fosfomycin-were detected in at least 95% of dominant-cluster genomes. The median antibiotic-class annotation hit count was 14 per genome (IQR, 13-15), including intrinsic and lineage-associated determinants. Predicted conjugative and mobilizable plasmid signatures occurred in 52.6% and 42.7% of dominant-cluster genomes, respectively. Across the full cohort, mobile-element and antibiotic-class hit counts were weakly positively correlated (Spearman ρ = 0.191). This unadjusted association was absent among 700 representatives selected one per ≤10-SNP component (ρ = -0.0721; p = 0.0567), and a component-bootstrap interval included zero. Assembly-adjusted associations differed between the full cohort and representatives. The small positive unadjusted collection-year association was not robust to location adjustment and was unsupported in dated representatives. Complete-genome annotations documented both chromosomal and plasmid blaOXA-23 copies and upstream ISAba1 proximity, without establishing promoter activity or resistance phenotype. These findings describe a cohesive ST2 lineage with heterogeneous accessory resistance and plasmid-mobility signatures, while emphasizing the sensitivity of genome-wide associations to assembly characteristics and clonal and geographic sampling.