Look for protein domains (identified by Chainsaw) that might have novel functional annotations
Take as input a directory full of predicted Chainsaw domains (in the form of pdb files) Extract the uniprot code for each pdb file Extract the residue numbers and sequence associated with the predicted protein domain Search for functional annotations on those domains
Data for leishmania infantum predicted domains can be found here: https://drive.google.com/drive/folders/1De5hrmzyd8o6gtX0rfKPpalXjEhJwGA3?usp=sharing
You can work out which residues (and their corresponding indices) are included in each predicted domain from the PDB file (where there are multiple PDB files for each protein - each one containing a predicted domain)
Residue indexes can be extracted from columns 23-26 in ATOM lines in the PDB file. Residue names can be extracted from columns 18-20 in ATOM lines.
Alternatively the Chainsaw predicted domains (written as residue indices) can be read from the csv files in chainsaw_domains_csv in the google drive.
From interpro: if name is 'uncharacterized' and go_terms='None' then this is one that we are interested in!