Machine learning reveals the selection pressure exerted by nonantibiotic pharmaceuticals at environmentally relevant concentrations on antibiotic resistance genotypes

Shaojing Sun,Yan Sun,Jialu Geng,Linlin Geng,Fan Meng,Qing Wang,Hong Qi
DOI: https://doi.org/10.1016/j.jenvman.2024.120829
Abstract:The emergence and increasing prevalence of antibiotic resistance pose a global public risk for human health, and nonantimicrobial pharmaceuticals play an important role in this process. Herein, five nonantimicrobial pharmaceuticals, including acetaminophen (ACT), clofibric acid (CA), carbamazepine (CBZ), caffeine (CF) and nicotine (NCT), tetracycline-resistant strains, five ARGs (sul1, sul2, tetG, tetM and tetW) and one integrase gene (intI1), were detected in 101 wastewater samples during two typical sewage treatment processes including anaerobic-oxic (A/O) and biological aerated filter (BAF) in Harbin, China. The impact of nonantibiotic pharmaceuticals at environmentally relevant concentrations on both the resistance genotypes and resistance phenotypes were explored. The results showed that a significant impact of nonantibiotic pharmaceuticals at environmentally relevant concentrations on tetracycline resistance genes encoding ribosomal protection proteins (RPPs) was found, while no changes in antibiotic phenotypes, such as minimal inhibitory concentrations (MICs), were observed. Machine learning was applied to further sort out the contribution of nonantibiotic pharmaceuticals at environmentally relevant concentrations to different ARG subtypes. The highest contribution and correlation were found at concentrations of 1400-1800 ng/L for NCT, 900-1500 ng/L for ACT and 7000-10,000 ng/L for CF for tetracycline resistance genes encoding RPPs, while no significant correlation was found between the target compounds and ARGs when their concentrations were lower than 500 ng/L for NCT, 100 ng/L for ACT and 1000 ng/L for CF, which were higher than the concentrations detected in effluent samples. Therefore, the removal of nonantibiotic pharmaceuticals in WWTPs can reduce their selection pressure for resistance genes in wastewater.
What problem does this paper attempt to address?