Tool to predict intrinsically disordered proteins.
pip install requirements.txt- For data generation: NCBI-BLAST+
- For making test/training sets: CD-HIT
hopfield_network.py: Hopfield network architecture implementation.idp_dataset.py: A script to generate embeddings for the dataset and load it into pytorch, along with all the associated augmentations.aaindex_feature_imporance.ipynb: A jupyter notebook where I experiment in order to get the most important biochemical features.basic_aa_features.csv: A csv file with the selected AAindex biochemical features for each AA.protein_feature_extractor.py: A python script to extract AAindex, evolutionary, and secondary structure features.make_training_set.py: A python script to generate the training/testing set from disprot.
I wrote a script that converts parquet files to csv files:
python3 ./parquet_to_csv.py path/to/parquet/file.parquet