TMscoring: Protein Structure Comparison Tool

TMscoring is a Python-based tool for comparing protein structures by calculating the TM-score and RMSD (Root Mean Square Deviation) values. It also provides functionalities for aligning protein structures and outputting aligned PDB files. the code has been developed by following TM_SCORE in Zhanglab.

Features

• Calculate TM-score: A measure of structural similarity between two protein structures.
• Calculate RMSD: A metric for the deviation between two aligned protein structures.
• Batch Processing: Process multiple structure files in specified directories.
• CSV Output: Save results (TM-score and RMSD) to a CSV file for easy analysis.
• Summary Statistics: Compute average TM-score and RMSD across all processed structures.

Requirements

• Python 3.7 or higher
• Required libraries:
  • numpy
  • scipy
  • biopython

To install the required packages, use:

pip install -r requirements.txt

Installation

Clone the Repository:

git clone https://github.com/farz1995/Tm_score.git
cd TM_score

python main.py

Usage

Example: Batch Processing with main.py This example compares protein structures in batch mode. The real structures are stored in real_structures/, and the predicted structures are in predicted_structures/.

Directory Structure real_structures/: Contains real protein structure files (e.g., 1a1b.pdb). predicted_structures/: Contains predicted protein structure files prefixed with structure_ (e.g., structure_1a1b.pdb).

Expected Output

CSV File: A file named result.csv containing the TM-score and RMSD for each protein pair. Summary Statistics: The average TM-score and RMSD are printed in the terminal.

Customization

You can edit the paths for real and predicted structure directories in main.py:

real_dir = "path/to/real_structures/"
pred_dir = "path/to/predicted_structures/"
output_csv = "result.csv"