High-risk antimicrobial resistance genomic traits in multidrug-resistant Salmonella across China's food chain: projections to 2050.

Antimicrobial resistance (AMR) is rising globally, yet factors shaping its genomic evolution remain poorly understood. We investigated how socioeconomic, climate, and health factors are associated with AMR spread in multi-drug resistant (MDR) Salmonella enterica across China's food chain, mapped hotspots and forecasted genomic trait prevalence to 2050. We analysed 2,536 MDR genomes from 15 provinces over 16 years, covering patients, food and non-food workers, meat, livestock, and environmental sources. A machine learning framework integrated genomic surveillance with risk assessment to identify high-risk AMR traits (clinically relevant, MDR, mobile). Regression models linked trait prevalence to socioeconomic, climate, health, and antimicrobial consumption data, informing projections to 2050. Eighty-four high-risk traits, including carbapenem, colistin, and β-lactam resistance, are projected to rise, with antibiotic use, population growth, healthcare spending, and urbanisation. Projections were trait-specific: blaCTX-M-64 and blaTEM-1 may increase up to 20%, while blaCTX-M-65 and blaCTX-M-14 remain below 5%. A mechanistic shift in fluoroquinolone resistance is projected: mutation-based resistance (gyrA 87) declines while plasmid-mediated genes (qnrB17, qnrS1) rise. Food workers emerged as a potential reservoir, showing higher trait prevalence than food sources, uniquely carrying last-resort antibiotic-resistance genes (blaTEM-30, blaNDM-5, blaCTX-M-64). Integrating genomic with socioeconomic surveillance could shift AMR interventions from reactive to preventive.