Global genomic surveillance of β-lactam resistance in Escherichia coli across human, animal, and environmental reservoirs.

BackgroundEscherichia coli poses a global health threat from increasing β-lactam resistance. This study uses genomic and One Health data to map resistance patterns and enhance antimicrobial resistance (AMR) prediction and management.MethodsThis study performed a One Health whole-genome analysis of 30,554 E. coli isolates from human, animal, and environmental sources, collected between 2000 and 2025. Publicly available genomic data were retrieved from NCBI, encompassing β-lactam resistance genes, encoding extended-spectrum β-lactamases, AmpC, and carbapenemases, along with key chromosomal resistance-associated mutations. Resistance determinants were identified and curated using AMRFinderPlus. Multilocus sequence typing (MLST) and pathotype assignments were conducted in silico. Temporal, host, sequence-type (ST), and geographic patterns of resistance were modeled, with gradient-boosted machine learning algorithms predicting minimum inhibitory concentrations (MICs) based on antibiotic resistance gene profiles and chromosomal features. Temporal and geographic resistance patterns were further analyzed using statistical and visualization tools in R and Python.ResultsA total of 30,554 E. coli whole-genome sequences from isolates spanning 126 countries between 2000 and 2025 were analyzed. The dataset included 2015 complete genomes, 177 chromosome-level assemblies, 8310 scaffold-level assemblies, and 20,052 contig-level assemblies, with an average sequence length of 5,120,983 bp. Isolates were categorized by source: 14,320 from humans, 8184 from animals, 3941 from environmental sources, and 4109 with an unknown origin. MLST was successfully performed for 29,648 isolates, identifying dominant STs such as ST131, ST11, and ST10. Human isolates were predominantly associated with epidemic clones ST131, ST73, and ST1193, while animal isolates were frequently associated with ST10, and environmental isolates showed a strong presence of ST155. Temporal analyses indicated a steady increase in β-lactam resistance, with blaCTX-M-15, blaNDM-5, and blaOXA-1 showing the strongest increases in prevalence. Model performance was strongest for carbapenems, particularly imipenem (R2 = 0.88) and ertapenem (R2 = 0.74), whereas predictions for ampicillin and piperacillin-tazobactam showed poor agreement with observed MICs.ConclusionsThis global analysis shows rising β-lactam resistance driven by key genes (blaCTX-M-15, blaNDM-5, blaOXA-1) in high-risk clones. Machine learning accurately predicted carbapenem MICs but struggled with β-lactam/β-lactamase inhibitor combinations. Integrating genomics within a One Health framework can strengthen AMR surveillance and control.