Comparing Machine Learning Algorithms for Identifying Antibiotic Resistance Genes (ARGs) in the Agricultural Soil Microbiome; Case Study in East Asia

Articles in Press

Document Type : Research Paper

Authors

1 Department of Soil Science, Faculty of Agriculture, Ferdowsi University of Mashhad, Mashhad, Iran

2 Soil Reclamation and Sustainable Land Management Research Department,, Soil and Water Research Institute of Iran, Agricultural Research, Education and Extension Organization (AREEO), Karaj, Iran

3 Department of Computer ,, Jihad Daneshgahi, of Khorasan Razavi, Mashhad, Iran

Abstract

With the escalating threat of antibiotic resistance, the accurate and comprehensive identification of antibiotic resistance genes (ARGs) in natural environments, particularly agricultural soils, has become a major concern in public and environmental health. In recent years, the application of machine learning algorithms has gained attention as a novel approach for analyzing complex metagenomic data and improving ARG detection. In this study, four machine learning algorithms—Random Forest, Gradient Boosting (XGBoost), Support Vector Machine (SVM), and Multilayer Perceptron (MLP)—were compared for their ability to identify resistance genes in the agricultural soil microbiome in India and China. Metagenomic data were obtained from the NCBI database and processed using the ARGs-OAP tool. A set of biological features, including GC content, amino acid frequency, and codon usage, was extracted. Statistical differences between resistant and non-resistant genes were assessed using the Mann–Whitney U test, and only significant features were selected for model training. The results demonstrated that the models, particularly Random Forest (with 98% accuracy), were capable of identifying resistance genes with high performance, even under conditions of imbalanced data and limited training sample size. These findings highlight the effectiveness of the selected biological features and machine learning algorithms in detecting ARGs in the agricultural soil microbiome in East Asia. This approach could serve as an efficient tool for environmental monitoring and policy-making aimed at controlling the spread of antibiotic resistance.

Keywords

Main Subjects

Iranian Journal of Soil and Water Research

Articles in Press, Accepted Manuscript
Available Online from 07 June 2026

Files

Share

How to cite

Statistics