def fetchPDBLigand(cci, filename=None): """Fetch PDB ligand data from PDB_ for chemical component *cci*. *cci* may be 3-letter chemical component identifier or a valid XML filename. If *filename* is given, XML file will be saved with that name. If you query ligand data frequently, you may configure ProDy to save XML files in your computer. Set ``ligand_xml_save`` option **True**, i.e. ``confProDy(ligand_xml_save=True)``. Compressed XML files will be save to ProDy package folder, e.g. :file:`/home/user/.prody/pdbligands`. Each file is around 5Kb when compressed. This function is compatible with PDBx/PDBML v 4.0. Ligand data is returned in a dictionary. Ligand coordinate atom data with *model* and *ideal* coordinate sets are also stored in this dictionary. Note that this dictionary will contain data that is present in the XML file and all Ligand Expo XML files do not contain every possible data field. So, it may be better if you use :meth:`dict.get` instead of indexing the dictionary, e.g. to retrieve formula weight (or relative molar mass) of the chemical component use ``data.get('formula_weight')`` instead of ``data['formula_weight']`` to avoid exceptions when this data field is not found in the XML file. URL and/or path of the XML file are returned in the dictionary with keys ``url`` and ``path``, respectively. Following example downloads data for ligand STI (a.k.a. Gleevec and Imatinib) and calculates RMSD between model (X-ray structure 1IEP) and ideal (energy minimized) coordinate sets: .. ipython:: python from prody import * ligand_data = fetchPDBLigand('STI') ligand_data['model_coordinates_db_code'] ligand_model = ligand_data['model'] ligand_ideal = ligand_data['ideal'] transformation = superpose(ligand_ideal.noh, ligand_model.noh) calcRMSD(ligand_ideal.noh, ligand_model.noh)""" if not isinstance(cci, str): raise TypeError('cci must be a string') if isfile(cci): inp = openFile(cci) xml = inp.read() inp.close() url = None path = cci cci = splitext(splitext(split(cci)[1])[0])[0].upper() elif len(cci) > 4 or not cci.isalnum(): raise ValueError('cci must be 3-letters long and alphanumeric or ' 'a valid filename') else: xml = None cci = cci.upper() if SETTINGS.get('ligand_xml_save'): folder = join(getPackagePath(), 'pdbligands') if not isdir(folder): makePath(folder) xmlgz = path = join(folder, cci + '.xml.gz') if isfile(xmlgz): with openFile(xmlgz) as inp: xml = inp.read() else: path = None #url = ('http://ligand-expo.rcsb.org/reports/{0[0]}/{0}/{0}' # '.xml'.format(cci.upper())) url = 'http://files.rcsb.org/ligands/download/{0}.xml'.format( cci.upper()) if not xml: #'http://www.pdb.org/pdb/files/ligand/{0}.xml' try: inp = openURL(url) except IOError: raise IOError( 'XML file for ligand {0} is not found online'.format(cci)) else: xml = inp.read() inp.close() if filename: out = openFile(filename, mode='w', folder=folder) out.write(xml) out.close() if SETTINGS.get('ligand_xml_save'): with openFile(xmlgz, 'w') as out: out.write(xml) import xml.etree.cElementTree as ET root = ET.XML(xml) if (root.get('{http://www.w3.org/2001/XMLSchema-instance}' 'schemaLocation') != 'http://pdbml.pdb.org/schema/pdbx-v40.xsd pdbx-v40.xsd'): LOGGER.warn('XML is not in PDBx/PDBML v 4.0 format, resulting ' 'dictionary may not contain all data fields') ns = root.tag[:root.tag.rfind('}') + 1] len_ns = len(ns) dict_ = {'url': url, 'path': path} for child in list(root.find(ns + 'chem_compCategory')[0]): tag = child.tag[len_ns:] if tag.startswith('pdbx_'): tag = tag[5:] dict_[tag] = child.text dict_['formula_weight'] = float(dict_.get('formula_weight')) identifiers_and_descriptors = [] results = root.find(ns + 'pdbx_chem_comp_identifierCategory') if results: identifiers_and_descriptors.extend(results) results = root.find(ns + 'pdbx_chem_comp_descriptorCategory') if results: identifiers_and_descriptors.extend(results) for child in identifiers_and_descriptors: program = child.get('program').replace(' ', '_') type_ = child.get('type').replace(' ', '_') dict_[program + '_' + type_] = child[0].text dict_[program + '_version'] = child.get('program_version') dict_['audits'] = [ (audit.get('action_type'), audit.get('date')) for audit in list(root.find(ns + 'pdbx_chem_comp_auditCategory')) ] atoms = list(root.find(ns + 'chem_comp_atomCategory')) n_atoms = len(atoms) ideal_coords = np.zeros((n_atoms, 3)) model_coords = np.zeros((n_atoms, 3)) atomnames = np.zeros(n_atoms, dtype=ATOMIC_FIELDS['name'].dtype) elements = np.zeros(n_atoms, dtype=ATOMIC_FIELDS['element'].dtype) resnames = np.zeros(n_atoms, dtype=ATOMIC_FIELDS['resname'].dtype) charges = np.zeros(n_atoms, dtype=ATOMIC_FIELDS['charge'].dtype) resnums = np.ones(n_atoms, dtype=ATOMIC_FIELDS['charge'].dtype) alternate_atomnames = np.zeros(n_atoms, dtype=ATOMIC_FIELDS['name'].dtype) leaving_atom_flags = np.zeros(n_atoms, np.bool) aromatic_flags = np.zeros(n_atoms, np.bool) stereo_configs = np.zeros(n_atoms, np.bool) ordinals = np.zeros(n_atoms, int) name2index = {} for i, atom in enumerate(atoms): data = dict([(child.tag[len_ns:], child.text) for child in list(atom)]) name = data.get('pdbx_component_atom_id', 'X') name2index[name] = i atomnames[i] = name elements[i] = data.get('type_symbol', 'X') resnames[i] = data.get('pdbx_component_comp_id', 'UNK') charges[i] = float(data.get('charge', 0)) alternate_atomnames[i] = data.get('alt_atom_id', 'X') leaving_atom_flags[i] = data.get('pdbx_leaving_atom_flag') == 'Y' aromatic_flags[i] = data.get('pdbx_atomatic_flag') == 'Y' stereo_configs[i] = data.get('pdbx_stereo_config') == 'Y' ordinals[i] = int(data.get('pdbx_ordinal', 0)) model_coords[i, 0] = float(data.get('model_Cartn_x', 0)) model_coords[i, 1] = float(data.get('model_Cartn_y', 0)) model_coords[i, 2] = float(data.get('model_Cartn_z', 0)) ideal_coords[i, 0] = float(data.get('pdbx_model_Cartn_x_ideal', 0)) ideal_coords[i, 1] = float(data.get('pdbx_model_Cartn_y_ideal', 0)) ideal_coords[i, 2] = float(data.get('pdbx_model_Cartn_z_ideal', 0)) pdbid = dict_.get('model_coordinates_db_code') if pdbid: model = AtomGroup(cci + ' model ({0})'.format(pdbid)) else: model = AtomGroup(cci + ' model') model.setCoords(model_coords) model.setNames(atomnames) model.setResnames(resnames) model.setResnums(resnums) model.setElements(elements) model.setCharges(charges) model.setFlags('leaving_atom_flags', leaving_atom_flags) model.setFlags('aromatic_flags', aromatic_flags) model.setFlags('stereo_configs', stereo_configs) model.setData('ordinals', ordinals) model.setData('alternate_atomnames', alternate_atomnames) dict_['model'] = model ideal = model.copy() ideal.setTitle(cci + ' ideal') ideal.setCoords(ideal_coords) dict_['ideal'] = ideal bonds = [] warned = set() for bond in list(root.find(ns + 'chem_comp_bondCategory') or bonds): name_1 = bond.get('atom_id_1') name_2 = bond.get('atom_id_2') try: bonds.append((name2index[name_1], name2index[name_2])) except KeyError: if name_1 not in warned and name_1 not in name2index: warned.add(name_1) LOGGER.warn('{0} specified {1} in bond category is not ' 'a valid atom name.'.format(repr(name_1), cci)) if name_2 not in warned and name_2 not in name2index: warned.add(name_2) LOGGER.warn('{0} specified {1} in bond category is not ' 'a valid atom name.'.format(repr(name_2), cci)) if bonds: bonds = np.array(bonds, int) model.setBonds(bonds) ideal.setBonds(bonds) return dict_
def parsePSF(filename, title=None, ag=None): """Returns an :class:`.AtomGroup` instance storing data parsed from X-PLOR format PSF file *filename*. Atom and bond information is parsed from the file. If *title* is not given, *filename* will be set as the title of the :class:`.AtomGroup` instance. An :class:`.AtomGroup` instance may be provided as *ag* argument. When provided, *ag* must have the same number of atoms in the same order as the file. Data from PSF file will be added to the *ag*. This may overwrite present data if it overlaps with PSF file content. This function now includes the angles, dihedrals, and impropers sections as well as donors, acceptors and crossterms!""" if ag is not None: if not isinstance(ag, AtomGroup): raise TypeError('ag must be an AtomGroup instance') psf = openFile(filename, 'rb') line = psf.readline() while line: line = line.strip() if line.endswith(b'!NATOM'): n_atoms = int(line.split(b'!')[0]) break line = psf.readline() if title is None: title = os.path.splitext(os.path.split(filename)[1])[0] else: title = str(title) if ag is None: ag = AtomGroup(title) else: if n_atoms != ag.numAtoms(): raise ValueError('ag and PSF file must have same number of atoms') serials = zeros(n_atoms, ATOMIC_FIELDS['serial'].dtype) segnames = zeros(n_atoms, ATOMIC_FIELDS['segment'].dtype) resnums = zeros(n_atoms, ATOMIC_FIELDS['resnum'].dtype) resnames = zeros(n_atoms, ATOMIC_FIELDS['resname'].dtype) atomnames = zeros(n_atoms, ATOMIC_FIELDS['name'].dtype) atomtypes = zeros(n_atoms, ATOMIC_FIELDS['type'].dtype) charges = zeros(n_atoms, ATOMIC_FIELDS['charge'].dtype) masses = zeros(n_atoms, ATOMIC_FIELDS['mass'].dtype) n = 0 n_bonds = 0 for line in psf: if line.strip() == b'': continue if b'!NBOND:' in line.upper(): items = line.split() n_bonds = int(items[0]) break if n + 1 > n_atoms: continue if len(line) <= 71: serials[n] = line[:8] segnames[n] = line[9:13].strip() resnums[n] = line[14:19] resnames[n] = line[19:23].strip() atomnames[n] = line[24:28].strip() atomtypes[n] = line[29:35].strip() charges[n] = line[35:44] masses[n] = line[50:60] else: items = line.split() serials[n] = items[0] segnames[n] = items[1] resnums[n] = items[2] resnames[n] = items[3] atomnames[n] = items[4] atomtypes[n] = items[5] charges[n] = items[6] masses[n] = items[7] n += 1 if n < n_atoms: raise IOError( 'number of lines in PSF atoms block is less than the number of ' 'atoms') n_angles = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NTHETA' in line: items = line.split() n_angles = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) b_array = fromstring(lines, count=n_bonds * 2, dtype=int, sep=' ') if len(b_array) != n_bonds * 2: raise IOError('number of bonds expected and parsed do not match') n_dihedrals = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NPHI' in line: items = line.split() n_dihedrals = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) a_array = fromstring(lines, count=n_angles * 3, dtype=int, sep=' ') if len(a_array) != n_angles * 3: raise IOError('number of angles expected and parsed do not match') n_impropers = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NIMPHI' in line: items = line.split() n_impropers = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) d_array = fromstring(lines, count=n_dihedrals * 4, dtype=int, sep=' ') if len(d_array) != n_dihedrals * 4: raise IOError('number of dihedrals expected and parsed do not match') n_donors = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NDON' in line: items = line.split() n_donors = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) i_array = fromstring(lines, count=n_impropers * 4, dtype=int, sep=' ') if len(i_array) != n_impropers * 4: raise IOError('number of impropers expected and parsed do not match') n_acceptors = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NACC' in line: items = line.split() n_acceptors = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) do_array = fromstring(lines, count=n_donors * 2, dtype=int, sep=' ') if len(do_array) != n_donors * 2: raise IOError('number of donors expected and parsed do not match') n_exclusions = 0 lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!NNB' in line: items = line.split() n_exclusions = int(items[0]) break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) ac_array = fromstring(lines, count=n_acceptors * 2, dtype=int, sep=' ') if len(ac_array) != n_acceptors * 2: raise IOError('number of acceptors expected and parsed do not match') lines = [] for i, line in enumerate(psf): if line.strip() == b'': continue if b'!' in line: break lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) nbe_array = fromstring(lines, count=n_exclusions * 2, dtype=int, sep=' ') if len(nbe_array) != n_exclusions * 2: raise IOError( 'number of nonbonded exclusions expected and parsed do not match') n_crossterms = 0 for i, line in enumerate(psf): if b'!NCRTERM' in line: items = line.split() n_crossterms = int(items[0]) break lines = [] for i, line in enumerate(psf): lines.append(line.decode(encoding='UTF-8')) lines = ''.join(lines) c_array = fromstring(lines, count=n_crossterms * 4, dtype=int, sep=' ') if len(c_array) != n_crossterms * 4: raise IOError('number of crossterms expected and parsed do not match') psf.close() ag.setSerials(serials) ag.setSegnames(segnames) ag.setResnums(resnums) ag.setResnames(resnames) ag.setNames(atomnames) ag.setTypes(atomtypes) ag.setCharges(charges) ag.setMasses(masses) if n_bonds > 0: b_array = add(b_array, -1, b_array) ag.setBonds(b_array.reshape((n_bonds, 2))) if n_angles > 0: a_array = add(a_array, -1, a_array) ag.setAngles(a_array.reshape((n_angles, 3))) if n_dihedrals > 0: d_array = add(d_array, -1, d_array) ag.setDihedrals(d_array.reshape((n_dihedrals, 4))) if n_impropers > 0: i_array = add(i_array, -1, i_array) ag.setImpropers(i_array.reshape((n_impropers, 4))) if n_donors > 0: do_array = add(do_array, -1, do_array) ag.setDonors(do_array.reshape((n_donors, 2))) if n_acceptors > 0: ac_array = add(ac_array, -1, ac_array) ag.setAcceptors(ac_array.reshape((n_acceptors, 2))) if n_exclusions > 0: nbe_array = add(nbe_array, -1, nbe_array) ag.setNBExclusions(nbe_array.reshape((n_exclusions, 2))) if n_crossterms > 0: c_array = add(c_array, -1, c_array) ag.setCrossterms(c_array.reshape((n_crossterms, 4))) return ag
def parsePSF(filename, title=None, ag=None): """Return an :class:`.AtomGroup` instance storing data parsed from X-PLOR format PSF file *filename*. Atom and bond information is parsed from the file. If *title* is not given, *filename* will be set as the title of the :class:`.AtomGroup` instance. An :class:`.AtomGroup` instance may be provided as *ag* argument. When provided, *ag* must have the same number of atoms in the same order as the file. Data from PSF file will be added to the *ag*. This may overwrite present data if it overlaps with PSF file content. Note that this function does not evaluate angles, dihedrals, and impropers sections.""" if ag is not None: if not isinstance(ag, AtomGroup): raise TypeError('ag must be an AtomGroup instance') psf = openFile(filename, 'rb') line = psf.readline() i_line = 1 while line: line = line.strip() if line.endswith('!NATOM'): n_atoms = int(line.split('!')[0]) break line = psf.readline() i_line += 1 if title is None: title = os.path.splitext(os.path.split(filename)[1])[0] else: title = str(title) if ag is None: ag = AtomGroup(title) else: if n_atoms != ag.numAtoms(): raise ValueError('ag and PSF file must have same number of atoms') serials = zeros(n_atoms, ATOMIC_FIELDS['serial'].dtype) segnames = zeros(n_atoms, ATOMIC_FIELDS['segment'].dtype) resnums = zeros(n_atoms, ATOMIC_FIELDS['resnum'].dtype) resnames = zeros(n_atoms, ATOMIC_FIELDS['resname'].dtype) atomnames = zeros(n_atoms, ATOMIC_FIELDS['name'].dtype) atomtypes = zeros(n_atoms, ATOMIC_FIELDS['type'].dtype) charges = zeros(n_atoms, ATOMIC_FIELDS['charge'].dtype) masses = zeros(n_atoms, ATOMIC_FIELDS['mass'].dtype) lines = psf.readlines(71 * (n_atoms + 5)) if len(lines) < n_atoms: raise IOError('number of lines in PSF is less than the number of ' 'atoms') for i, line in enumerate(lines): if i == n_atoms: break i_line += 1 if len(line) <= 71: serials[i] = line[:8] segnames[i] = line[9:13].strip() resnums[i] = line[14:19] resnames[i] = line[19:23].strip() atomnames[i] = line[24:28].strip() atomtypes[i] = line[29:35].strip() charges[i] = line[35:44] masses[i] = line[50:60] else: items = line.split() serials[i] = items[0] segnames[i] = items[1] resnums[i] = items[2] resnames[i] = items[3] atomnames[i] = items[4] atomtypes[i] = items[5] charges[i] = items[6] masses[i] = items[7] i = n_atoms while 1: line = lines[i].split() if len(line) >= 2 and line[1] == '!NBOND:': n_bonds = int(line[0]) break i += 1 lines = ''.join(lines[i + 1:]) + psf.read(n_bonds / 4 * 71) array = fromstring(lines, count=n_bonds * 2, dtype=int, sep=' ') if len(array) != n_bonds * 2: raise IOError('number of bonds expected and parsed do not match') psf.close() ag.setSerials(serials) ag.setSegnames(segnames) ag.setResnums(resnums) ag.setResnames(resnames) ag.setNames(atomnames) ag.setTypes(atomtypes) ag.setCharges(charges) ag.setMasses(masses) array = add(array, -1, array) ag.setBonds(array.reshape((n_bonds, 2))) return ag