This script downloads a number of random protein structures with a given structural classification and writes the relevant protein chain into a mmCIF file.

download_protein_family.py

from os.path import join, isdir
from os import mkdir
import numpy as np
import biotite.structure as struc
import biotite.structure.io.pdb as pdb
import biotite.structure.io.pdbx as pdbx
import biotite.database.rcsb as rcsb


# The structural classifier
SCOP_ID = "1000000"
# The number of samples
N = 10
# The directory for output CIF files
OUT_DIR = "output"


### Find suitable structures
classification_db_query = rcsb.FieldQuery(
    "rcsb_polymer_instance_annotation.type",
    exact_match="SCOP2"
)
scop_id_query = rcsb.FieldQuery(
    "rcsb_polymer_instance_annotation.annotation_lineage.id",
    exact_match=SCOP_ID
)
full_query = classification_db_query & scop_id_query
# Not only search for PDB entries, bit for specific protein chains
# with the given classification
pdb_chains = rcsb.search(full_query, return_type="polymer_instance")
# Ensure reproducible behavior
np.random.seed(0)
selected_pdb_chains = np.random.choice(sorted(pdb_chains), N)


### Fetch selected structures and write relevant chains to hard drive
if not isdir(OUT_DIR):
    mkdir(OUT_DIR)
for pdb_chain in selected_pdb_chains:
    pdb_id, chain = pdb_chain.split(".")
    # Use PDBx format to get 'official' chain IDs rather than the ones
    # provided by the authors to correctly match the chain value from
    # the search query
    mmtf_file = pdbx.PDBxFile.read(rcsb.fetch(pdb_id, "pdbx"))
    structure = pdbx.get_structure(mmtf_file, model=1, use_author_fields=False)
    # Filter relevant protein chain
    chain = structure[
        struc.filter_amino_acids(structure) & (structure.chain_id == chain)
    ]
    # Write filtered chain to output CIF file
    pdbx_file = pdbx.PDBxFile()
    pdbx.set_structure(pdbx_file, chain, data_block=pdb_id)
    pdbx_file.write(join(OUT_DIR, f"{pdb_id}.cif"))