Parsing a .pdb file in Python

I am mildly surprised that no one mentioned the Bio.PDB package from BioPython. Writing a PDB parser on your own is a rather serious case of unnecessarily reinventing, I mean reimplementing the wheel.

BioPython is a useful collection of packages for working with other kinds of biological data (e.g. protein or nucleic acid sequences) as well.

Update

I added a few functions I use for downloading and reading PDB files by way of example:

import os import sys import urllib.request import Bio import Bio.PDB import Bio.SeqRecord def download_read_pdb(pdbcode, datadir, keepfile=True): """ Downloads a PDB file from the Internet and saves it in a data directory. Then it reads and returns the structure inside. :param pdbcode: The standard PDB ID e.g. '3ICB' :param datadir: The directory where the downloaded file will be saved :param keepfile: if False, then the downloaded file will be deleted (default: keep the downloaded file) :return: a Bio.PDB Structure object or None if something went wrong """ pdbfilenm = download_pdb(pdbcode, datadir) if pdbfilenm is None: return None struct = read_pdb(pdbcode, pdbfilenm) if not keepfile: os.remove(pdbfilenm) return struct def download_pdb(pdbcode, datadir, downloadurl="http://files.rcsb.org/download/"): """ Downloads a PDB file from the Internet and saves it in a data directory. :param pdbcode: The standard PDB ID e.g. '3ICB' or '3icb' :param datadir: The directory where the downloaded file will be saved :param downloadurl: The base PDB download URL, cf. `https://www.rcsb.org/pages/download/http#structures` for details Note that the unencrypted HTTP protocol is used by default to avoid spurious OpenSSL errors... :return: the full path to the downloaded PDB file or None if something went wrong """ pdbfn = pdbcode + ".pdb" url = downloadurl + pdbfn outfnm = os.path.join(datadir, pdbfn) try: urllib.request.urlretrieve(url, outfnm) return outfnm except Exception as err: # all sorts of things could have gone wrong... print(str(err), file=sys.stderr) return None def read_pdb(pdbcode, pdbfilenm): """ Read a PDB structure from a file. :param pdbcode: A PDB ID string :param pdbfilenm: The PDB file :return: a Bio.PDB.Structure object or None if something went wrong """ try: pdbparser = Bio.PDB.PDBParser(QUIET=True) # suppress PDBConstructionWarning struct = pdbparser.get_structure(pdbcode, pdbfilenm) return struct except Exception as err: print(str(err), file=sys.stderr) return None def extract_seqrecords(pdbcode, struct): """ Extracts the sequence records from a Bio.PDB structure. :param pdbcode: the PDB ID of the structure, needed to add a sequence ID to the result :param struct: a Bio.PDB.Structure object :return: a list of Bio.SeqRecord objects """ ppb = Bio.PDB.PPBuilder() seqrecords = [] for i, chain in enumerate(struct.get_chains()): # extract and store sequences as list of SeqRecord objects pps = ppb.build_peptides(chain) # polypeptides seq = pps[0].get_sequence() # just take the first, hope there's no chain break seqid = pdbcode + chain.id seqrec = Bio.SeqRecord.SeqRecord(seq, id=seqid, description="Sequence #{}, {}".format(i+1, seqid)) seqrecords.append(seqrec) return seqrecords def get_calphas(struct): """ Extracts the C-alpha atoms from a PDB structure. :param struct: A Bio.PDB.Structure object. :return: A list of Bio.PDB.Atom objects representing the C-alpha atoms in `struct`. """ calphas = [ atom for atom in struct.get_atoms() if atom.get_fullname() == " CA " ] return calphas 

I am aware that this was an old question but maybe someone still finds this pointer helpful.

It looks like a library to parse PDB files already exists. Check out:

http://prody.csb.pitt.edu/

Briefly looking through the tutorials, it seems like this is what you're talking about no?

http://www.csb.pitt.edu/prody/tutorial.html#atomic-data

For large pdb files with a lot of atoms, there is no blank left between fields, so you cannot use the split command. Instead, you can use the Protein Data Bank format definition to parse the pdb file:

with open('min.pdb') as pdbfile: for line in pdbfile: if line[:4] == 'ATOM' or line[:6] == "HETATM": print line # Split the line splitted_line = [line[:6], line[6:11], line[12:16], line[17:20], line[21], line[22:26], line[30:38], line[38:46], line[46:54]] print splitted_line # To format again the pdb file with the fields extracted print "%-6s%5s %4s %3s %s%4s %8s%8s%8s\n"%tuple(splitted_line) 

Below is a sample output:

# Original line: HETATM10000 O HOH B3257 2.509 40.006 -4.097 1.00 0.00 O # Fields extracted: ['HETATM', '10000', ' O ', 'HOH', 'B', '3257', ' 2.509', ' 40.006', ' -4.097'] # Reformatted line: HETATM10000 O HOH B3257 2.509 40.006 -4.097