Performance And Mechanism Of Multi-Soil-Layering System For The Removal Of Antibiotic Resistance Genes From Rural Sewage

With the development of the rural economy and the improvement of the living standard of the population,the production of rural sewage gradually increased.The issue of the bacterial antibiotic resistance in sewage has been a concern at national and international scale.At the present stage,the rural sewage treatment technology in China still remains the problem of high operating costs,complex maintenance and poor adaptability,and how to solve the bacterial antibiotic-resistant contamination in the sewage has become a challenge.Currently,multi-soil-layering system（MSL）had been gradually emphasised as a low-cost and easy-to-maintain rural sewage treatment technology,however,there was a lack of research on the performance and mechanism of the MSL system on the disposal of antibiotic resistance genes（ARGs）in rural sewage.Hence,this study was conducted to optimise the effect of MSL system on rural sewage treatment,with emphasis on the removal of ARGs from rural sewage under different scenarios of hydraulic loads（200,400,600 L/m²/d）,sewage pollution loads（original concentration,75%dilution,50%dilution）,and influent patterns（intermittent and continuous influent）,to identify the critical factors that drive the MSL system to eliminate the ARGs from rural sewage,and to reveal the major mechanism of the system in the efficient removal of ARGs from the sewage.Further,an efficiency prediction model for MSL system to treat the ARG of sewage was constructed based on a machine learning algorithm,and screened the pivotal feature vectors which affect the prediction performance.It aims to provide the theoretical foundation and technical support to further improve the steady-state efficient treatment and utilisation of rural wastewater.The main results obtained in this study are as follows:（1）Under different scenarios,the MSL system was able to remove ARGs from the effluent effectively,especially for polymyxins,aminoglycosides,multi-drug-resistant,and tetracyclines ARGs,with an average removal rate ranging from 49.60%to 61.03%.The best removal efficacy of the MSL system for effluent ARGs was achieved when the system was under the scenario conditions of 400L/m²/d hydraulic load,original effluent pollution load,and continuous influent,in which the MSL system was operated stably against temperature fluctuations,and the removal efficiencies of the ARGs did not differ significantly between the warm and cold seasons（p>0.05）,with removal rates ranging from 64.19%to 80.05%.（2）The structural model revealed that the microbial composition of sewage played a dominant role in the treatment of ARGs by the MSL system,while the physicochemical factors of sewage and the system matrix mainly enhanced the removal of sewage ARGs by the system through indirect effects,and the key physicochemical factors were EC,Cl^-,p H,and NO₃-N etc.The MSL system can also remove sewage ARGs by the microbial composition of sewage and the substrate.In addition,network analysis and random forest model were used to refine the features of microbial community changes in the wastewater treatment process,and the core microorganisms in the sewage and matrix affecting the ARGs of sewage treated by the MSL system such as Fluvicola,Pseudomonas,and Arthrobacter were clarified.（3）The removal of ARGs from sewage by the MSL system was a synergistic effect of physical adsorption,chemical inactivation and biocompetitive processes.The matrix was able to effectively adsorb and capture host bacteria carrying ARGs and pathogenic microorganisms carrying mobile elements in the effluent,thereby reducing the presence and proliferation of drug-resistant bacteria and ARGs.Subsequently,it was found that the presence of functional groups,such as-OH,within the system was able to eliminate the host bacteria adsorbed by the matrix,and the reshaping of the microbial community of the system inhibited the colonisation and proliferation of ARGs-carrying host bacteria,which led to the effective removal of sewage ARGs.（4）Based on the KEGG database comparison analysis,it was found that the functional genes of microbial C and N metabolism in the matrix of the MSL system were able to synergistically eliminate effluent ARG.In particular,the effects of matrix SMB on ARG capture and abatement were mainly facilitated by the enhancement of functional genes of the N pathway of N metabolism,such as ammonia oxidation,nitrate reduction,and nitrite ammonification.In contrast,the effect of substrate PL microorganisms on ARG capture and abatement mainly functioned through the C metabolism organic carbon oxidation pathway.（5）Application of the comparative support vector machine,random forest,and gradient boosting regression models identified that the latter two were able to accurately predict the efficacy of treating wastewater ARG removal for the MSL system,with R² of 0.796 and 0.795,respectively.Of these,the gradient boosting model was the optimal model for prediction,with prediction accuracy R² of 0.915after optimisation with the Bayesian algorithm.And effluent EC,ARG subtype,ARG type,and effluent p H were the important features that affect the model prediction of effluent ARG removal efficacy of MSL system for treating effluent.