RF-TICA-MD

This repository contains the codes and scripts for the paper "Resolving Protein Conformational Plasticity and Substrate Binding Through the Lens of Machine-Learning" by Navjeet Ahalawat and Jagannath Mondal (2022) doi: https://doi.org/10.1101/2022.01.07.475334

I am using Anaconda (Python 3.6) and the following packages:

• MSMBuilder 3.8.0.
• PyEMMA 2.5.4
• scikit-learn 0.19.2
• numpy
• matplotlib

Files

• run_pair_distances.sh - bash script to calculate distances between given pairs using msmbuilder package
• contact_pairs.txt - contact pairs used in in msmbuilder
• script_randomforest.py - python script to build the Random Forest Classifier Model
• xgboost.py - python script to build XGBoost model for classification
• selected_pairs2npy.py - Python script to extract the selected pairs identified from RF Classifier
• Top200_residue_pairs - Top 200 pairs from the RF Classifier Model of T4L system
• plot_tica.py - Python script to plot the free energy surface
• plot_tica_fig2.py - Python script to plot the free energy surface
• plot_tica_fig2_new.py - Python script to plot the free energy surface
• plot_tica_fig4.py - Python script to plot the free energy surface
• plot_tica_fig4_new.py - Python script to plot the free energy surface
• traj1_pairs_1-2.tcl - VMD TCL script to visulaize network of the selected pairs
• traj2_pairs_1-2.tcl - VMD TCL script to visulaize network of the selected pairs
• traj3_pairs_1-2.tcl - VMD TCL script to visulaize network of the selected pairs
• t4l_image.com - UCSF Chimera script to visulaize the T4L protein

preparing data and adapting script

1. Prepare raw collective varibale using msmbuilder script.
2. Catenate all the numpy files and rename it to X_3_trajs.npy
3. Calculate distance between cavity COM and ligand (benzene) COM
4. Catenate all distance files in the same order of X_3_trajs.npy and convert it into binary format 0 and 1 using distance cutoff criteria. Rename this file as Y_3_trajs.npy.
5. Build RF Classifier Model using script_randomforrest.py
6. Use this model for further analysis

Generating Network Images using vmd, ref.pdb, and tcl scripts

$ vmd -e traj1_pairs_1-2.tcl ref.pdb

$ vmd -e traj2_pairs_1-2.tcl ref.pdb

$ vmd -e traj3_pairs_1-2.tcl ref.pdb