TWN-Align: A Topological Water Network–Guided Molecular Alignment Method for Fragment Screening

This repository provides the source code for TWN-Align, a fragment screening and alignment framework guided by Topological Water Networks (TWNs). The overall workflow of TWN-Align is summarized below.

This code has been tested on Windows using Python 3.9.

A YAML file containing all required dependencies is provided. You can set up a Conda virtual environment as follows:

conda env create -f TWN-FS-open-source.yaml
conda activate TWN-FS-open-source

To run TWN-Align, the following preset files are required:

1. trajectory
2. TWN-Region
3. TWN-Pattern
4. TWN-pharma
5. Fragment files

The trajectory, TWN-Region and TWN-Pattern files can be generated using the same procedure as in the TWN-FG method:

https://github.com/RgJeoung/TWN-FG

Important notes on fragment files Fragment files must contain energy-minimized 3D structures. Hydrogen atoms must be explicitly included to enable pharmacophore feature assignment.

Details on generating the TWN-pharma file are provided below.

The TWN-pharma file is generated by computing the dipole moments of TWNs. It can be obtained using the following command:

python TWN_dipole_extractor.py -d ./DATA/TWN-Pattern

After preparing all required preset files, you can run the main script, TWN-FS-open-source.py. Please ensure that all input directories are organized in the same manner as the example DATA directory provided in this repository.

You can check the available command-line options using:

python TWN-FS-open-source.py -h

This will display the following instructions:

-h, --help            show this help message and exit
-f FRAGMENT, --fragment FRAGMENT
                      Path to the fragment directory (input fragments to be screened)
-r REGION, --region REGION
                      Region definition file used for TWN analysis
-o OUTPUT, --output OUTPUT
                      Output directory name (optional).
                      If not specified, a directory named:
                        <fragment_dir>_TWN-FS-RESULT0001
                      will be created automatically.
                      If the directory already exists, the index will be incremented.
-a True/False, --alignment True/False
                      Enable TWN-based alignment step.
                      Default: True
-c CORE, --core CORE  Number of CPU cores to use for parallel processing.
                      Default: maximum available cores

To run TWN-Align with default settings, use the following command:

python TWN-FS-open-source.py -f ./DATA/test_fragments -r ./DATA/TWN-Region/CHK1_1NVR_R4/TWN_Region_1.sdf

After completion, an output directory such as the following will be created:

test_fragments_TWN-FS-RESULT0001

The output directory contains:

1. Aligned fragment files (.sdf)
2. TWN-FS.log (log and report file)

We recommend visualizing the aligned fragment files together with the corresponding TWN-Region to verify that the alignment was performed correctly.

The TWN-FS.log file includes the execution settings as well as a brief screening report.

For a detailed interpretation of the output files and methodology, please refer to our published paper.

Questions Please contact our professor: 📧 nskang@cnu.ac.kr

Bugs / Feature Requests Please submit a GitHub issue or contact the developer directly: 📧 ray971125@naver.com