How abdollahegazy/comp_dock works
Project type and purpose
• This is a command-line, Python-based data-processing pipeline (no UI) composed of standalone scripts.
• Purpose: automate ligand–protein complex processing (AF3 and MD), detect errors, select viable poses, merge complexes, compute RMSD for downstream analysis.
File structure
• Top level contains four main processing scripts:
• find_empty_ligand.py (error detection)
• most_viable.py (pose viability ranking)
• merger.py (complex merging)
• rmsd_test.py (RMSD calculation)
• No frontend, no database, no services—pure file-based I/O.
• README describes inputs (AF3- and MD-derived PDBQT/PDB files) and outputs (lists of empty ligands, best poses, merged complexes, RMSD metrics).
Key architectural patterns and design principles
• Pipeline / batch-processing architecture: each script is a stage.
• Modularity: each stage has a single responsibility.
• Data-flow orientation: outputs of one stage feed into the next.
• File-based interfaces: standardized filenames and directories passed via CLI args or hard-coded paths.
Main components
A. Input Data Sources
– AF3-derived docking results (PDBQT files)
– MD simulation outputs (PDB/PDBQT)
B. Processing Modules (pipeline stages)
1. Error Detector (find_empty_ligand.py)
• Reads docking results
• Outputs empty_ligands.txt
2. Pose Viability Analyzer (most_viable.py)
• Reads original outputs + empty_ligands.txt
• Ranks ligand poses, outputs viable_pose_list.txt
3. Complex Merger (merger.py)
• Reads viable poses + processed receptor structures
• Produces merged complexes (PDB files)
4. RMSD Calculator (rmsd_test.py)
• Reads merged complexes + AF3 reference
• Computes per-atom RMSD, outputs RMSD_report.csv
C. Outputs / Artifacts
– error lists, ranked pose lists, merged complexes, RMSD reports
D. External Libraries
– Python 3.x
– Potential dependencies: BioPython or MDAnalysis, numpy, pandas
Detail without overthinking
• Four scripts as separate components.
• Key file artifact.
• Sequencing and dependencies.