Predicting peptide binding

This package uses an API from PepSite2, which "can predict binding of a given peptide onto a protein structure".

The API provided by PepSite is limited to only one peptide sequence. This package allows the researcher to give a full protein sequence, split it into a list of peptide sequences, returning the predictions for each sequence.

The researcher also has the ability to choose the peptide sequence size, as well as the number of binding scores returned per prediction.

This package gives added functionality to the base PepSite, but to use results, please cite the original paper:
PepSite: prediction of peptide-binding sites from protein surfaces.
Trabuco LG, Lise S, Petsalaki E, and Russell RB.
Nucleic Acids Res. 2012; 40(Web Server issue):W423-426.

Usage

run predict_peptide_binding, passing the following arguments:

- peptide_src_type: if the peptide is a file or a string.
- peptide: provide path (if file) or the peptide string.
- pdb_code: such as 1eak.
- pdb_chain: such as D.
- window_length: splits the peptide sequence into 
                    subsequences of size window_length. 
                    Required if peptide_src_type is a file 
                    (-f).
- [score_count]: optional, returns the top scores 
                    rather than all scores of the 
                    prediction. Default is 10. 

Currently, the sliding window only works when providing the peptide sequence in file format, and the sequence length must be greater than the window_length provided.

Examples

Running with a peptide string:

python3 -m predict_peptide_binding -s GPAGPPGA 1eak D 10 3

Running with a peptide sequence file, ace2.txt:

python3 -m predict_peptide_binding -f ace2.txt 1eak D 10 3