It has been demonstrated that antibiotic resistance genes (ARGs) exhibit seasonal variations in municipal wastewater treatment plants (MWTPs), but their relationship to bacterial phylogeny structure remains unclear. Using advanced metagenomic techniques and machine learning approach, the current study conducted a year-long investigation to explore the relationship between ARGs and the bacterial community of activated sludge in a full-scale MWTP in Beijing, where seasonal dynamics are remarkable. High abundance of ARGs, notably the clinically relevant high-risk ARGs, was observed in winter and spring, the cold season in Beijing. Seasonal patterns were also observed in the diversity of ARGs and the overall bacterial community. Machine learning-based random forest classification models were utilized to identify biomarkers for ARGs and bacterial genera as indicators of seasonal differences. Subsequent analysis of the relationship between ARGs and bacterial biomarkers was examined using random forest regression models. Results showed that the enrichment of potential pathogens such as Mycobacterium, Clostridium and Pseudomonas was high in winter and spring, strongly contributing to the abundance of high-risk ARGs (ermB, aac(6′)-I, tetM, blaTEM, and mefA) during cold season. Conversely, functional taxa associated with activated sludge, such as Thauera, displayed seasonal fluctuations and a preference for ARGs with minor clinical implications. Metagenomic binning further illustrated the contribution of Mycobacterium to ARG enrichment in cold season. Our findings highlight the collective impact of human-derived clinically relevant taxa and functional bacterial taxa in activated sludge on the seasonal dynamics of ARGs in MWTPs. Additionally, this study offers valuable insights into the safe disposal of the excess sludge from MWTPs.
Read full abstract