bogdanmagometa / protein-rmsd-calculation

Calculate Root Mean Square Deviation between native and predicted strcutures of peptide. Visualize complexes

protein-rmsd-calculation

Project structure

• calc_rmsd.py - Add rmsd1, rmsd2, rmsd3 columns to the metrics.csv table, save aligned filtered PDBs.
• make_movies.py - Create movies from aligned filtered PDBs.
• add_text_all.py - Overlay text on all movies.
• arrange_grid.sh - Arrange all movies into one grid of specified grid width and height.

Prerequisites

• Linux OS
• Maybe macOS (at least finish_launching might not work)
• You need to have conda installed
• To generate videos without watermarks, license might be needed. Educational license is free: https://pymol.org/edu/

Environment setup

$ conda create -n rmsd_calculation python=3.10
$ conda activate rmsd_calculation
$ conda install numpy pandas
$ conda install -c conda-forge prody    # used in calc_rmsd.py only
$ conda install -c conda-forge -c schrodinger pymol-bundle  # used in make_movies.py only
$ conda install -c conda-forge ffmpeg   # used in add_text_all.py, add_text.sh

Usage example

Assuming the following directory tree structure:

./
    output/
        data/
            structures/
                ...directories each containing true and predicted structures...
            metrics.csv
    datasets/
        npz_data/
            ...npz-files...

$ python3 calc_rmsd.py --data_dir ./output/data --npz_data ./datasets/npz_data
$ python3 make_movies.py --structures_dir ./output/data/structures --videos_dir ./videos
$ python3 add_text_all.py --metrics_path ./output/data/metrics_with_rmsds.csv --videos_dir ./videos --videos_with_text_dir ./videos_with_text
$ ./arrange_grid.sh ./videos_with_text ./all_structs_video.mp4 5 8