NanoBEP – A Machine Learning Based Tool for Nanobody Binding Energy Prediction (Dataset)

This repository contains codes, data for machine learning models developed to predict the binding affinity of various protein complexes. The repository explores the interactions of Antigen-Antibody (Ag-Ab), Protein-Protein (P-P), and Nanobody-Protein (Nb-P) complexes through different experimental setups and modeling approaches.

Repository Structure

Codes:

Contains the codes used to build the models for Set A, Set B, Set C using Random forest and the codes for support vector regression analysis.

Data:

Contains the dataset for training and testing the models.

Feature Generation:

Contains the codes and data to generate AAPP features, PPDX features and SASA, Prodigy features.

Graphical_elimination:

Contains scripts and data for removing redundant features using network analysis.

Requirements

• Python: 3.8+

• Libraries:

     o   scikit-learn

     o   numpy

     o   pandas

     o   matplotlib

     o   scipy

     o   joblib

Usage

Run the Jupyter Notebook scripts located in the respective folders to train and test the models.