Hierarchical approach

We tried different machine learning models (SVM, MLP, KNN, LR, RF), and features extracted by ESM2, to predict different types of nucleic acid binding proteins using hierarchical approach, including the non-NABP/NABP classification, DBP/RBP classification and SSB/DSB classification. Files in the folder 1-non-NABP-NABP are the pre-trained models to predict non-NABPs and NABPs, and files in the folder 2-DBP-RBP are the pre-trained models to predict DBPs and RBPs. Each folder contains the README.txt file, indicating the user note for each model. For the SSBs and DSBs prediction, see https://figshare.com/articles/software/Prediction_of_SSB_DSB/25927360.<br>