alphaFilter is a Python script that processes a directory of AlphaFold2 models and outputs a list of models based on a mean confidence threshold. This is especially useful for filtering out AlphaFold2 models below a certain confidence level and ensuring that the models have a minimum number of desired residues. The goal of the script is for the user to quickly find high-confidence alphafold models without pulling models created from fragmentary sequences.
- Filter AlphaFold2 models by their plddt confidence scores.
- Ensure models have a minimum number of residues.
- Simple and easy-to-use command-line interface.
- Python 3.6 or later
- AlphaFold2 models in a directory
To use alphaFilter, simply provide the path to the directory containing the AlphaFold2 models, the desired plddt confidence threshold, and the minimum number of residues a model should have.
python3 alphafilter.py -d DIRECTORY -t THRESHOLD -r RES_MINIM-d DIRECTORY, --directory DIRECTORY: Path to the directory containing the AlphaFold2 models.-t THRESHOLD, --threshold THRESHOLD: Threshold for plddt confidence. The default value is 90.-r RES_MINIM, --res_minim RES_MINIM: Filter minimum number of residues in AlphaFold2 structure.
-h, --help: Shows the help message and exits.
For a directory of AlphaFold2 models located at ./models, to filter out models with a plddt confidence score below 92 and with less than 100 residues:
python3 alphafilter.py -d ./models -t 92 -r 100Feel free to open issues or submit pull requests if you find any bugs or have suggestions for improvements.