PROPTIMUS PRIME

PROPTIMUS PRIME: Per-residue optimisation of protein structures: Primary integrity measures

PROPTIMUS PRIME is a workflow providing repair of amino-acid side chains in protein structures predicted by the AlphaFold2 algorithm. It can be either run locally as a command line application or integrated as a regular Python library. It uses the following libraries: Biopython, RDKit, scikit-learn, and the PDB2PQR suite.

Command line use

Setup

git clone https://github.com/sb-ncbr/proptimus_prime
cd proptimus_prime/
python -m venv venv_prime
source venv_prime/bin/activate
pip install -r requirements.txt

Execution

One file

python prime.py <input> [output] [-l log] [-d -s]

Batch run

This runs PRIME over all structures in a folder, using 10 CPU cores. This was used for the analysis of Swiss-Prot-proteins predictions in AlphaFold DB.

python executor_prime.py <input_dir> 

input: path to the PDB file to process (mandatory)
output: path to the corrected file (optional)
log: path to the log file (optional, must be preceded by -l)
-d: delete auxilliary files (also deletes the log file, if its path has not been specified)
-s: silent mode input_dir: path to the directory with PDB files (mandatory)

Python library integration

Import the PrimaryIntegrityMeasuresTaker into your Python script. All options listed in Command line use are available within the constructor parameters. The logger parameter is used for processing large sets of PDB files using a custom script. Ignore for casual use.