DSSP Toolkit

DSSP Toolkit is a suite designed to run and compare multiple DSSP (Define Secondary Structure of Proteins) implementations simultaneously. Because tools like Biopython, OST, and PDB-REDO often vary in speed and assignment logic, this toolkit helps you choose the best implementation for your specific structural biology workflow.

---

✨ Features

• Multi-Tool Support: Compare mkdssp, OpenStructure (OST), Biopython, and PDB-REDO side-by-side.
• Performance Tracking: Detailed timing metrics for each implementation.
• Comparison Engine: Identifies discrepancies in secondary structure assignments at the residue level.
• Parallel Processing: Optimized with Python's multiprocessing for handling large datasets of .cif files.
• Resilient Design: Modular "Methods" architecture; if one tool fails, the others continue.

---

🏗️ Architecture

Component	Description
main.py	Orchestrator. Manages CLI arguments, process pooling, and report generation.
methods/	Isolated Python modules for each DSSP implementation.
scripts/	Custom Bash wrappers for mkdssp and raw CSV parsing (Experimental).
data/	Default directory for input .cif structure files.

---

Quick Start

Run with defaults (processes data/test_pdb/ in parallel):

python main.py

Command Reference

Flag	Argument	Description
--pdb-dir	<path>	Directory containing .cif files (Default: data/test_pdb).
--pdb-file	<path>	Path to a single specific .cif file.
--tools	<list>	Select specific tools (Choices: ost, pdbredo, biopython, bash).
--no-parallel	(Flag)	Disable multiprocessing (run files one by one).
--summary	(Flag)	Automatically run statistical analysis after benchmarking.

Examples

1. Run on a custom directory with analysis:

python main.py --pdb-dir /home/data/structures --summary

2. Benchmark specific tools (OST vs PDB-REDO) on a single file:

python main.py --pdb-file test.cif --tools ost pdbredo

3. Debugging mode (Serial execution, single tool):

python main.py --pdb-file test.cif --tools ost --no-parallel

🛠️ Installation

The toolkit includes an automated setup script to handle dependencies, including mkdssp and the conda environment.If you have not installed the tools for dssp or having a hard time doing that you can

# Clone the repository
git clone [https://github.com/bhatmohsin/DSSP-Toolkit)
cd dssp-toolkit

# Run the automated installer
chmod +x install.sh
./install.sh

# Activate the environment
conda activate dssp_benchmark