Abstract
Introduction: Hospital environments are important reservoirs of antimicrobial resistance (AMR), where microbial communities, antibiotic selection pressure, and mobile genetic elements facilitate the emergence and transmission of resistance traits. Objective: The study aimed to characterize the hospital microbiota, resistance structure, and transmission dynamics using integrated metagenomics and network analysis. Materials and Methods: Multicenter longitudinal observational study in five tertiary-care hospitals (January - December 2025). In total, 150 samples were obtained from intensive care units, surgical wards, operating theaters, sink drains, wastewater outlets, ventilation systems, and high-touch surfaces. After quality control, 143 samples were sequenced by whole metagenome on the Illumina NovaSeq 6000 platform. Top-notch bioinformatics pipelines were used for taxonomic profiling, (ARG) detection, mobile genetic element analysis, microbial network construction, and predictive modeling. Results: Pseudomonas (18.4%), Acinetobacter (15.1%), and Klebsiella (12.6%) represented the genera most commonly isolated. Wastewater and sink drains had the highest microbial diversity and ARG burden. The most common were β-lactam resistance genes (28.6%), while clinically important determinants included bla_KPC (41.2%), bla_NDM (38.6%), mecA (35.1%), and mcr-1 (12.4%) In summary, 71.8% of ARGs were related to mobile genetic elements. ROA modelling with a Random Forest accurately identified major ARG reservoirs (AUROC = 0.94), and three novel resistance gene variants were detected. Conclusion: Hospital wastewater, sink drains, and ICU-related areas are recognized as high-density reservoirs and spreaders of antimicrobial resistance. This study integrates metagenomic surveillance and predictive analytics to identify high-risk reservoirs and support targeted infection control strategies.