How to retrieve PPI interface, protein-ligand interface and protein-DNA/RNA interface?

[yang-arina]

Hi! Is the PISA API working?
I currently have a series of unp_ids, such as ['Q5VTD9', 'Q6PL24', 'Q9UKM9', 'Q9H0A3', 'Q9P2N6', 'P21127', 'Q8TD94', 'Q86SH2', 'Q9BYR3', 'Q99856']. How can I retrieve for each unp_id:

• 1. all protein complexes and their PPI interface residues;
• 2. all protein-ligand complexes and their protein interface residues;
• 3. all protein-DNA/RNA complexes and their protein interface residues.

If you could help me with this, I would be extremely grateful!

[NatureGeorge]

i. and iii can be fulfill with current version of pdb-profiling.

# homomeric 
pdb_profiling sifts-mapping --input demo_uniprot.txt --column UniProt --func pipe_select_ho --output he_result.txt
# heteromeric
pdb_profiling sifts-mapping --input demo_uniprot.txt --column UniProt --func pipe_select_he --output he_result.txt
# protein-DNA/RNA
pdb_profiling sifts-mapping --input demo_uniprot.txt --column UniProt --func pipe_select_else --kwargs 'func="Protein/NA"' --output na_result.txt

The PISA API have been fixed but also been updated, requiring some modification with the code and scripts. (https://www.ebi.ac.uk/pdbe/news/improved-macromolecular-interactions-data-pisa-lite)
I would try to implement the new PISA API in the following weeks.

At the mean time, you can try another tool with similar functionalities: https://github.com/ELELAB/PDBminer (https://pubs.acs.org/doi/10.1021/acs.jcim.3c00884). But their tool also seems to have some bugs to fix.

[NatureGeorge]

Actually, to save your time and effort, you can directly turn to https://www.pdbbind-plus.org.cn/ for retrieving such interacting data.

EDIT: It seems that they do not provide the details of interacting residues.

[yang-arina]

Actually, to save your time and effort, you can directly turn to https://www.pdbbind-plus.org.cn/ for retrieving such interacting data.

EDIT: It seems that they do not provide the details of interacting residues.

Thank u for your kind reply! I'll try through PISA API to get the interface residue!

[NatureGeorge]

Actually, to save your time and effort, you can directly turn to https://www.pdbbind-plus.org.cn/ for retrieving such interacting data.
EDIT: It seems that they do not provide the details of interacting residues.

Thank u for your kind reply! I'll try through PISA API to get the interface residue!

The valid PISA API list is here https://www.ebi.ac.uk/pdbe/api/pisa.html

[yang-arina]

Hi, I would like to inquire about how to calculate the BS score. It's used to measure the mapping quality between UniProt and PDB chains, with a maximum value of 1. I have the UniProt fasta file and the corresponding PDB chain's pdb structure file. Could you please advise me on how to proceed with the calculation of the BS score? Thank you for your assistance.

[NatureGeorge]

Sorry for the late reply. I have been busy around these days. The functionalities of BS score caculation and representative structure selection process for interacting proteins seems need to be restored. I came across https://pypi.org/project/arctic3d/ (https://www.nature.com/articles/s42003-023-05718-w) and found it useful. You can try it first.

[NatureGeorge]

Oh. Maybe I misunderstood your needs. The way to calculate the BS score is straightforward. It composed of aligned, CN-terminal, Insertion, and Deletion scoring parts. I recommand you to only calculate the aligned part and normalized it with the full length of UniProt if you want to compute it by your self, it is enough for most cases.

BS score is intended for selecting representative strcuctures of each UniProt. The calculated results can be found at the bs_score column (or bs_score_1 and bs_score_2 for protein 1 and protein 2 respectively in ho/he/na_results.txt) of result.txt. According to the BS score, as well the the exact coverage region the the PDB to UniProt, we can select those PDB structures with as less coverage overlap but more overall coverage of the UniProt full length as possible. The selection boolean indicator can be found at the select_tag column (or select_tag_1, select_tag_2).

For interacting pairs, we focus on the different coverage of the interface residues to the UniProt full length pair, the selection boolean indicator can be found at the i_select_tag column.