def buildFromChain(self, chain):
"""Build from a :class:`.Chain`."""
assert isinstance(chain, Chain), 'chain must be a Chain instance'
gaps = self._gaps
residues = list(chain.iterResidues())
temp = residues[0].getResnum()-1
protein_resnames = set(getKeywordResnames('protein'))
for res in chain:
if not res.getResname() in protein_resnames:
continue
resid = res.getResnum()
incod = res.getIcode()
aa = AAMAP.get(res.getResname(), 'X')
simpres = SimpleResidue(resid, aa, incod, res)
if gaps:
diff = resid - temp - 1
if diff > 0:
self._seq += NONE_A * diff
temp = resid
self._seq += aa
self._list.append(simpres)
self._dict[(resid, incod)] = simpres
self._title = 'Chain {0:s} from {1:s}'.format(chain.getChid(),
chain.getAtomGroup().getTitle())
def _parsePDBLines(atomgroup, lines, split, model, chain, subset,
altloc_torf, format='PDB'):
"""Return an AtomGroup. See also :func:`.parsePDBStream()`.
:arg lines: PDB/PQR lines
:arg split: starting index for coordinate data lines"""
format = format.upper()
if format == 'PDB':
isPDB = True
else:
isPDB = False
if subset is not None:
subset = subset.lower()
if subset in ('calpha', 'ca'):
subset = set(('CA',))
elif subset in ('backbone', 'bb'):
subset = set(getBackboneAtomNames())
only_subset = True
protein_resnames = set(getKeywordResnames('protein'))
else:
only_subset = False
if chain is None:
only_chains = False
else:
only_chains = True
onlycoords = False
n_atoms = atomgroup.numAtoms()
if n_atoms > 0:
asize = n_atoms
else:
# most PDB files contain less than 99999 atoms
asize = min(len(lines) - split, 99999)
addcoords = False
if atomgroup.numCoordsets() > 0:
addcoords = True
alength = asize
coordinates = np.zeros((asize, 3), dtype=float)
atomnames = np.zeros(asize, dtype=ATOMIC_FIELDS['name'].dtype)
resnames = np.zeros(asize, dtype=ATOMIC_FIELDS['resname'].dtype)
resnums = np.zeros(asize, dtype=ATOMIC_FIELDS['resnum'].dtype)
chainids = np.zeros(asize, dtype=ATOMIC_FIELDS['chain'].dtype)
hetero = np.zeros(asize, dtype=ATOMIC_FIELDS['hetero'].dtype)
altlocs = np.zeros(asize, dtype=ATOMIC_FIELDS['altloc'].dtype)
icodes = np.zeros(asize, dtype=ATOMIC_FIELDS['icode'].dtype)
serials = np.zeros(asize, dtype=ATOMIC_FIELDS['serial'].dtype)
if isPDB:
segnames = np.zeros(asize, dtype=ATOMIC_FIELDS['segment'].dtype)
elements = np.zeros(asize, dtype=ATOMIC_FIELDS['element'].dtype)
bfactors = np.zeros(asize, dtype=ATOMIC_FIELDS['beta'].dtype)
occupancies = np.zeros(asize, dtype=ATOMIC_FIELDS['occupancy'].dtype)
secondary = None
anisou = None
siguij = None
else:
charges = np.zeros(asize, dtype=ATOMIC_FIELDS['charge'].dtype)
radii = np.zeros(asize, dtype=ATOMIC_FIELDS['radius'].dtype)
asize = 2000 # increase array length by this much when needed
start = split
stop = len(lines)
nmodel = 0
# if a specific model is requested, skip lines until that one
if isPDB and model is not None and model != 1:
for i in range(split, len(lines)):
if lines[i][:5] == 'MODEL':
nmodel += 1
if model == nmodel:
start = i+1
stop = len(lines)
break
if nmodel != model:
raise PDBParseError('model {0:d} is not found'.format(model))
if isinstance(altloc_torf, str):
if altloc_torf.strip() != 'A':
LOGGER.info('Parsing alternate locations {0:s}.'
.format(altloc_torf))
which_altlocs = ' ' + ''.join(altloc_torf.split())
else:
which_altlocs = ' A'
altloc_torf = False
else:
which_altlocs = ' A'
altloc_torf = True
acount = 0
altloc = defaultdict(list)
i = start
END = False
while i < stop:
line = lines[i]
startswith = line[0:6]
if startswith == 'ATOM ' or startswith == 'HETATM':
if only_subset:
atomname = line[12:16].strip()
resname = line[17:21].strip()
if not (atomname in subset and resname in protein_resnames):
i += 1
continue
else:
atomname = line[12:16]
resname = line[17:21]
chid = line[21]
if only_chains:
if not chid in chain:
i += 1
continue
alt = line[16]
if alt not in which_altlocs:
altloc[alt].append((line, i))
i += 1
continue
try:
coordinates[acount, 0] = line[30:38]
coordinates[acount, 1] = line[38:46]
coordinates[acount, 2] = line[46:54]
except:
if acount >= n_atoms > 0:
if nmodel ==0:
raise ValueError(format + 'file and AtomGroup ag must '
'have same number of atoms')
LOGGER.warn('Discarding model {0:d}, which contains more '
'atoms than first model does.'.format(nmodel+1))
acount = 0
nmodel += 1
coordinates = np.zeros((n_atoms, 3), dtype=float)
while lines[i][:6] != 'ENDMDL':
i += 1
else:
raise PDBParseError('invalid or missing coordinate(s) at '
'line {0:d}.'.format(i+1))
if onlycoords:
acount += 1
i += 1
continue
serials[acount] = line[6:11]
altlocs[acount] = alt
atomnames[acount] = atomname
resnames[acount] = resname
chainids[acount] = chid
resnums[acount] = line[22:26]#.split()[0])
icodes[acount] = line[26]
if isPDB:
try:
occupancies[acount] = line[54:60]
except:
LOGGER.warn('failed to parse occupancy at line {0:d}'
.format(i))
try:
bfactors[acount] = line[60:66]
except:
LOGGER.warn('failed to parse beta-factor at line {0:d}'
.format(i))
hetero[acount] = startswith[0] == 'H'
segnames[acount] = line[72:76]
elements[acount] = line[76:78]
else:
try:
charges[acount] = line[54:62]
except:
LOGGER.warn('failed to parse charge at line {0:d}'
.format(i))
try:
radii[acount] = line[62:69]
except:
LOGGER.warn('failed to parse radius at line {0:d}'
.format(i))
acount += 1
if n_atoms == 0 and acount >= alength:
# if arrays are short extend them with zeros
alength += asize
coordinates = np.concatenate(
(coordinates, np.zeros((asize, 3), float)))
atomnames = np.concatenate((atomnames,
np.zeros(asize, ATOMIC_FIELDS['name'].dtype)))
resnames = np.concatenate((resnames,
np.zeros(asize, ATOMIC_FIELDS['resname'].dtype)))
resnums = np.concatenate((resnums,
np.zeros(asize, ATOMIC_FIELDS['resnum'].dtype)))
chainids = np.concatenate((chainids,
np.zeros(asize, ATOMIC_FIELDS['chain'].dtype)))
hetero = np.concatenate((hetero,
np.zeros(asize, ATOMIC_FIELDS['hetero'].dtype)))
altlocs = np.concatenate((altlocs,
np.zeros(asize, ATOMIC_FIELDS['altloc'].dtype)))
icodes = np.concatenate((icodes,
np.zeros(asize, ATOMIC_FIELDS['icode'].dtype)))
serials = np.concatenate((serials,
np.zeros(asize, ATOMIC_FIELDS['serial'].dtype)))
if isPDB:
bfactors = np.concatenate((bfactors,
np.zeros(asize, ATOMIC_FIELDS['beta'].dtype)))
occupancies = np.concatenate((occupancies,
np.zeros(asize, ATOMIC_FIELDS['occupancy'].dtype)))
segnames = np.concatenate((segnames,
np.zeros(asize, ATOMIC_FIELDS['segment'].dtype)))
elements = np.concatenate((elements,
np.zeros(asize, ATOMIC_FIELDS['element'].dtype)))
if anisou is not None:
anisou = np.concatenate((anisou, np.zeros((asize, 6),
ATOMIC_FIELDS['anisou'].dtype)))
if siguij is not None:
siguij = np.concatenate((siguij, np.zeros((asize, 6),
ATOMIC_FIELDS['siguij'].dtype)))
else:
charges = np.concatenate((charges,
np.zeros(asize, ATOMIC_FIELDS['charge'].dtype)))
radii = np.concatenate((radii,
np.zeros(asize, ATOMIC_FIELDS['radius'].dtype)))
#elif startswith == 'END ' or startswith == 'CONECT':
# i += 1
# break
elif startswith == 'ENDMDL' or startswith[:3] == 'END':
if acount == 0:
# If there is no atom record between ENDMDL & END skip to next
i += 1
continue
if model is not None:
i += 1
break
diff = stop - i - 1
if diff < acount:
END = True
if onlycoords:
if acount < n_atoms:
LOGGER.warn('Discarding model {0:d}, which contains '
'{1:d} fewer atoms than the first model '
'does.'.format(nmodel+1, n_atoms-acount))
else:
coordsets[nmodel] = coordinates
nmodel += 1
acount = 0
if not END:
coordinates = coordsets[nmodel]
else:
if acount != n_atoms > 0:
raise ValueError('PDB file and AtomGroup ag must have '
'same number of atoms')
# this is where to decide if more coordsets should be expected
if END:
if addcoords:
atomgroup.addCoordset(coordinates[:acount])
else:
atomgroup._setCoords(coordinates[:acount])
else:
coordsets = np.zeros((diff/acount+1, acount, 3))
coordsets[0] = coordinates[:acount]
onlycoords = True
if not only_subset:
atomnames = np.char.strip(atomnames[:acount])
resnames = np.char.strip(resnames[:acount])
atomgroup.setNames(atomnames[:acount])
atomgroup.setResnames(resnames[:acount])
atomgroup.setResnums(resnums[:acount])
atomgroup.setChids(chainids[:acount])
atomgroup.setHeteros(hetero[:acount])
atomgroup.setAltlocs(altlocs[:acount])
atomgroup.setIcodes(np.char.strip(icodes[:acount]))
atomgroup.setSerials(serials[:acount])
if isPDB:
atomgroup.setBetas(bfactors[:acount])
atomgroup.setOccupancies(occupancies[:acount])
atomgroup.setSegnames(np.char.strip(segnames[:acount]))
atomgroup.setElements(np.char.strip(elements[:acount]))
if anisou is not None:
atomgroup.setAnisous(anisou[:acount] / 10000)
if siguij is not None:
atomgroup.setAnistds(siguij[:acount] / 10000)
else:
atomgroup.setCharges(charges[:acount])
atomgroup.setRadii(radii[:acount])
nmodel += 1
n_atoms = acount
acount = 0
coordinates = np.zeros((n_atoms, 3), dtype=float)
if altloc and altloc_torf:
_evalAltlocs(atomgroup, altloc, chainids, resnums,
resnames, atomnames)
altloc = defaultdict(list)
if END:
break
elif isPDB and startswith == 'ANISOU':
if anisou is None:
anisou = True
anisou = np.zeros((alength, 6),
dtype=ATOMIC_FIELDS['anisou'].dtype)
try:
index = acount - 1
anisou[index, 0] = line[28:35]
anisou[index, 1] = line[35:42]
anisou[index, 2] = line[43:49]
anisou[index, 3] = line[49:56]
anisou[index, 4] = line[56:63]
anisou[index, 5] = line[63:70]
except:
LOGGER.warn('failed to parse anisotropic temperature '
'factors at line {0:d}'.format(i))
elif isPDB and startswith =='SIGUIJ':
if siguij is None:
siguij = np.zeros((alength, 6),
dtype=ATOMIC_FIELDS['siguij'].dtype)
try:
index = acount - 1
siguij[index, 0] = line[28:35]
siguij[index, 1] = line[35:42]
siguij[index, 2] = line[43:49]
siguij[index, 3] = line[49:56]
siguij[index, 4] = line[56:63]
siguij[index, 5] = line[63:70]
except:
LOGGER.warn('failed to parse standard deviations of '
'anisotropic temperature factors at line {0:d}'.format(i))
elif startswith =='SIGATM':
pass
i += 1
if onlycoords:
if acount == atomgroup.numAtoms():
coordsets[nmodel] = coordinates
nmodel += 1
if nmodel == coordsets.shape[0]:
if addcoords:
atomgroup.addCoordset(coordsets)
else:
atomgroup._setCoords(coordsets)
else:
if addcoords:
atomgroup.addCoordset(coordsets[:nmodel])
else:
atomgroup._setCoords(coordsets[:nmodel])
elif not END:
# this means last line wast an ATOM line, so atomgroup is not decorated
if addcoords:
atomgroup.addCoordset(coordinates[:acount])
else:
atomgroup._setCoords(coordinates[:acount])
if not only_subset:
atomnames = np.char.strip(atomnames[:acount])
resnames = np.char.strip(resnames[:acount])
atomgroup.setNames(atomnames[:acount])
atomgroup.setResnames(resnames[:acount])
atomgroup.setResnums(resnums[:acount])
atomgroup.setChids(chainids[:acount])
atomgroup.setHeteros(hetero[:acount])
atomgroup.setAltlocs(altlocs[:acount])
atomgroup.setIcodes(np.char.strip(icodes[:acount]))
atomgroup.setSerials(serials[:acount])
if isPDB:
if anisou is not None:
atomgroup.setAnisous(anisou[:acount] / 10000)
if siguij is not None:
atomgroup.setAnistds(siguij[:acount] / 10000)
atomgroup.setSegnames(np.char.strip(segnames[:acount]))
atomgroup.setElements(np.char.strip(elements[:acount]))
atomgroup.setBetas(bfactors[:acount])
atomgroup.setOccupancies(occupancies[:acount])
else:
atomgroup.setCharges(charges[:acount])
atomgroup.setRadii(radii[:acount])
if altloc and altloc_torf:
_evalAltlocs(atomgroup, altloc, chainids, resnums, resnames, atomnames)
return atomgroup
