TWN-FS (TWN-based Fragment Screening)

A novel fragment screening method for drug discovery

Hye Ree Yoon, Gyoung Jin Park, Anand Balupuri, Nam Sook Kang *

Computational and Structural Biotechnology Journal 21 (2023) 4683–4696

Introduction

Water molecules contribute to structure and function of proteins and also influence the environment within the binding pocket

Water molecules form a variety of hydrogen-bonded cyclic water-ring networks, collectively known as topological water networks (TWNs)

Analyzing the TWNs in protein binding sites can provide valuable insights into potential locations and shapes for fragments within the binding site

Here, we introduce TWN-based fragment screening (TWN-FS) method, a novel screening method that suggests fragments through grouped TWN analysis within the protein binding site

Installation

This code was tested in windows with Python 3.9

A yaml file containing all requirements is provided

This can be readily setup using conda

conda env create -f TWN-env.yaml
conda activate TWN-env

Standard Usage

TWN_automation.py

The fastest way to run TWN-FS

Analyze TWNs-fragment of multiple proteins in one running

TWN_analysis.py and TWN_gridbox.py are operated sequentially

TWNs of cyclin dependent kinase 2 (CDK2) and fragments of subregion are provided as example data

Provided file

Files that can be changed according to user analysis

1. TWNs (data/TWNs_CDK2/.pdb file)
2. KinFragLib AP region fragment (data/region/AP.mol2)
3. reference_form (data/reference_form.xlsx)

Calculate programs and files

• shaep.exe (data/ShaEP): calculates shape similarity
• TWN_automation.py: analyze and summarize grouped TWNs (includes TWN_gridbox.py and TWN_analysis.py)
• TWN_gridbox.py: generate grouped TWNs
• TWN_analysis.py: analyze and summarize grouped TWNs–fragment shape similarity and average distance

Protein coordinate information

• Protein_CDK2.pdb (data/Protein_CDK2.pdb) Klifs coordinate, CDK2 PDB 4ERW)

Input file :

• Protein TWNs extracted after MD simulation (data/TWNs_CDK2/.pdb file)
• AP region fragment (data/region/AP.mol2): AP (.mol2) file is KinFragLib's AP site fragments
• reference_form (data/reference_form.xlsx): includes fragment details (ID, Kinase, PDB, SMILES, region). reference_form (.xlsx) file is KinFragLib's AP site fragments information

Output file :

• TWN.pdb : pdb file (grouped TWNs-fragment visualization, only oxygen atoms are displayed)
• Summary.xlsx : summary of results grouped TWNs-fragment calculation scores (shape similarity and average distance)
• TWN_Group_inform.xlsx : information(TWNs-configuration, coordinates, counts) of grouped TWNs-fragment
• TWN_Group_Centor_Point.tsv : centroids of grouped TWNs-fragment
• OBMol : sdf file converted from pdb file
• Distance_Average, ShaEP : calculation scores (shape similarity and average distance)
• Analysis.txt, TWN_gridbox.txt : log files of .py that was executed

python TWN_automation.py -twn parent parent_path_of_twn_folder -region subregion_path -shaep shaep_path -ref reference_file -out output_path 

For example:

python TWN_automation.py -twn ./data/TWNs -region ./data/region -shaep ./data/ShaEP/shaep.exe -ref ./data/reference_form.xlsx -out ./data/output

Advanced Usage

TWN_gridbox.py

Grouping TWNs-fragment (primary clustering and secondary clustering)

Define the protein file to display in Complex.pdb ('-pt' option)

Define the range of grouped TWNs-fragments to extract

The criteria is ligand file ('-bd' option)

Create a complex.pdb that visualizes the protein together so that the location of the grouped TWNs-fragments is known ('-c' option)

Output files :

• Complex.pdb : visualize grouped TWNs-fragment and protein structure
• TWN.pdb
• TWN_gridbox.txt
• TWN_Group_Centor_Point.tsv

python TWN_gridbox.py -twn TWN_file -pt protein_file -o output_path -c complex 

For example:

python TWN_gridbox.py -twn ./data/TWNs/TWNs_CDK2 -pt ./data/Protein_CDK2.pdb -o ./data/output -c 

TWN_anlysis.py

Calculate shape similarity and distance average as a result of TWN_gridbox.py

Sets the criteria for selecting shape similarity and distance average for grouped TWNs-fragments, ascending or descending ('-cond' option)

Output files :

• same as TWN_automation.py

python TWN_anlysis.py -twn path_of_twn_folder -region subregion_path -shaep shaep_path -ref reference_file -out output_path -cond set_condition_of_extract_values

For example:

python TWN_anlysis.py -twn ./data/TWNs/TWNs_CDK2 -region ./data/region -shaep ./data/ShaEP/shaep.exe -ref ./data/reference_form.xlsx -out ./data/output -cond True

Contact (Questions/Bugs/Requests)

Questions : Please ask our professor nskang@cnu.ac.kr

Bugs/Requests : Please submit a GitHub issue or contact me rudwls2717@naver.com

Acknowledgements

Thank you for our Laboratory https://homepage.cnu.ac.kr/cim/index.do

If you find this code useful, please consider citing our work