Fig. 2

Architecture of the proposed model to predict arginine methylated sites. a Data collection, removal of redundant sequences, and data balancing using under-sampling. b Feature extraction including the dipeptide composition, physicochemical properties, amino acid composition, and information theory-based features. c Build a model for different machine learning algorithms. d Model evaluation using various evaluation parameters based on 10-fold cross-validation. e Model interpretation of model outputs using SHAP algorithm