def _fillMolecule(name, resname, chain, resid, insertion, coords, segid,
element, occupancy, beta, charge, record):
numAtoms = len(name)
mol = Molecule()
mol.empty(numAtoms)
mol.name = np.array(name, dtype=mol._dtypes['name'])
mol.resname = np.array(resname, dtype=mol._dtypes['resname'])
mol.chain = np.array(chain, dtype=mol._dtypes['chain'])
mol.resid = np.array(resid, dtype=mol._dtypes['resid'])
mol.insertion = np.array(insertion, dtype=mol._dtypes['insertion'])
mol.coords = np.array(np.atleast_3d(np.vstack(coords)),
dtype=mol._dtypes['coords'])
mol.segid = np.array(segid, dtype=mol._dtypes['segid'])
mol.element = np.array(element, dtype=mol._dtypes['element'])
mol.occupancy = np.array(occupancy, dtype=mol._dtypes['occupancy'])
mol.beta = np.array(beta, dtype=mol._dtypes['beta'])
# mol.charge = np.array(charge, dtype=mol._dtypes['charge'])
# mol.record = np.array(record, dtype=mol._dtypes['record'])
return mol
def _applyProteinCaps(mol, caps):
# AMBER capping
# =============
# This is the (horrible) way of adding caps in tleap:
# For now, this is hardwired for ACE and NME
# 1. Change one of the hydrogens of the N terminal (H[T]?[123]) to the ACE C atom, giving it a new resid
# 1a. If no hydrogen present, create the ACE C atom.
# 2. Change one of the oxygens of the C terminal ({O,OT1,OXT}) to the NME N atom, giving it a new resid
# 2a. If no oxygen present, create the NME N atom.
# 3. Reorder to put the new atoms first and last
# 4. Remove the lingering hydrogens of the N terminal and oxygens of the C terminal.
# Define the atoms to be replaced (0 and 1 corresponds to N- and C-terminal caps)
terminalatoms = {
'ACE': ['H1', 'H2', 'H3', 'HT1', 'HT2', 'HT3'],
'NME': ['OXT', 'OT1', 'O']
} # XPLOR names for H[123] and OXT are HT[123]
# and OT1, respectively.
capresname = ['ACE', 'NME']
capatomtype = ['C', 'N']
# For each caps definition
for seg in caps:
prot = mol.atomselect(
'protein'
) # Can't move this out since we remove atoms in this loop
# Get the segment
segment = np.where(mol.segid == seg)[0]
# Test segment
if len(segment) == 0:
raise RuntimeError(
'There is no segment {} in the molecule.'.format(seg))
if not np.any(prot & (mol.segid == seg)):
raise RuntimeError(
'Segment {} is not protein. Capping for non-protein segments is not supported.'
.format(seg))
# For each cap
passed = False
for i, cap in enumerate(caps[seg]):
if cap is None or (isinstance(cap, str) and cap == 'none'):
continue
# Get info on segment and its terminals
segidm = mol.segid == seg # Mask for segid
segididx = np.where(segidm)[0]
resids = mol.resid[segididx]
terminalids = [segididx[0], segididx[-1]]
terminalresids = [resids[0], resids[-1]]
residm = mol.resid == terminalresids[i] # Mask for resid
if not passed:
orig_terminalresids = terminalresids
passed = True
if cap is None or cap == '': # In case there is no cap defined
logger.warning(
'No cap provided for resid {} on segment {}. Did not apply it.'
.format(terminalresids[i], seg))
continue
elif cap not in capresname: # If it is defined, test if supported
raise RuntimeError(
'In segment {}, the {} cap is not supported. Try using {} instead.'
.format(seg, cap, capresname))
# Test if cap is already applied
testcap = np.where(segidm & residm & (mol.resname == cap))[0]
if len(testcap) != 0:
logger.warning(
'Cap {} already exists on segment {}. Did not re-apply it.'
.format(cap, seg))
continue
# Test if the atom to change exists
termatomsids = np.zeros(residm.shape, dtype=bool)
for atm in terminalatoms[cap]:
termatomsids |= mol.name == atm
termatomsids = np.where(termatomsids & segidm & residm)[0]
if len(termatomsids) == 0:
# Create new atom
termcaid = np.where(segidm & residm & (mol.name == 'CA'))[0]
termcenterid = np.where(segidm & residm
& (mol.name == capatomtype[1 - i]))[0]
atom = Molecule()
atom.empty(1)
atom.record = np.array(['ATOM'],
dtype=Molecule._dtypes['record'])
atom.name = np.array([capatomtype[i]],
dtype=Molecule._dtypes['name'])
atom.resid = np.array([terminalresids[i] - 1 + 2 * i],
dtype=Molecule._dtypes['resid'])
atom.resname = np.array([cap],
dtype=Molecule._dtypes['resname'])
atom.segid = np.array([seg], dtype=Molecule._dtypes['segid'])
atom.element = np.array([capatomtype[i]],
dtype=Molecule._dtypes['element'])
atom.chain = np.array(
[np.unique(mol.chain[segidm])],
dtype=Molecule._dtypes['chain']
) # TODO: Assumption of single chain in a segment might be wrong
atom.coords = mol.coords[termcenterid] + 0.33 * np.subtract(
mol.coords[termcenterid], mol.coords[termcaid])
mol.insert(atom, terminalids[i])
# logger.info('In segment {}, resid {} had none of these atoms: {}. Capping was performed by creating '
# 'a new atom for cap construction by tleap.'.format(seg, terminalresids[i],
# ' '.join(terminalatoms[cap])))
else:
# Select atom to change, do changes to cap, and change resid
newatom = np.max(termatomsids)
mol.set('resname', cap, sel=newatom)
mol.set('name', capatomtype[i], sel=newatom)
mol.set('element', capatomtype[i], sel=newatom)
mol.set('resid', terminalresids[i] - 1 + 2 * i,
sel=newatom) # if i=0 => resid-1; i=1 => resid+1
# Reorder
neworder = np.arange(mol.numAtoms)
neworder[newatom] = terminalids[i]
neworder[terminalids[i]] = newatom
_reorderMol(mol, neworder)
# For each cap
for i, cap in enumerate(caps[seg]):
if cap is None or (isinstance(cap, str) and cap == 'none'):
continue
# Remove lingering hydrogens or oxygens in terminals
mol.remove('segid {} and resid "{}" and name {}'.format(
seg, orig_terminalresids[i], ' '.join(terminalatoms[cap])),
_logger=False)
# Remove terminal hydrogens regardless of caps or no caps. tleap confuses itself when it makes residues into terminals.
# For example HID has an atom H which in NHID should become H[123] and crashes tleap.
for seg in np.unique(mol.get('segid', 'protein')):
segidm = mol.segid == seg # Mask for segid
segididx = np.where(segidm)[0]
resids = mol.resid[segididx]
mol.remove('(resid "{}" "{}") and segid {} and hydrogen'.format(
resids[0], resids[-1], seg),
_logger=False)
def _applyProteinCaps(mol, caps):
# AMBER capping
# =============
# This is the (horrible) way of adding caps in tleap:
# For now, this is hardwired for ACE and NME
# 1. Change one of the hydrogens of the N terminal (H[T]?[123]) to the ACE C atom, giving it a new resid
# 1a. If no hydrogen present, create the ACE C atom.
# 2. Change one of the oxygens of the C terminal ({O,OT1,OXT}) to the NME N atom, giving it a new resid
# 2a. If no oxygen present, create the NME N atom.
# 3. Reorder to put the new atoms first and last
# 4. Remove the lingering hydrogens of the N terminal and oxygens of the C terminal.
# Define the atoms to be replaced (0 and 1 corresponds to N- and C-terminal caps)
terminalatoms = {
"ACE": ["H1", "H2", "H3", "HT1", "HT2", "HT3"],
"NME": ["OXT", "OT1", "O"],
} # XPLOR names for H[123] and OXT are HT[123]
# and OT1, respectively.
capresname = ["ACE", "NME"]
capatomtype = ["C", "N"]
# For each caps definition
for seg in caps:
prot = mol.atomselect(
"protein"
) # Can't move this out since we remove atoms in this loop
# Get the segment
segment = np.where(mol.segid == seg)[0]
# Test segment
if len(segment) == 0:
raise RuntimeError(f"There is no segment {seg} in the molecule.")
if not np.any(prot & (mol.segid == seg)):
raise RuntimeError(
f"Segment {seg} is not protein. Capping for non-protein segments is not supported."
)
# For each cap
passed = False
for i, cap in enumerate(caps[seg]):
if cap is None or (isinstance(cap, str) and cap == "none"):
continue
# Get info on segment and its terminals
segidm = mol.segid == seg # Mask for segid
segididx = np.where(segidm)[0]
resids = mol.resid[segididx]
terminalids = [segididx[0], segididx[-1]]
terminalresids = [resids[0], resids[-1]]
residm = mol.resid == terminalresids[i] # Mask for resid
if not passed:
orig_terminalresids = terminalresids
passed = True
if cap is None or cap == "": # In case there is no cap defined
logger.warning(
f"No cap provided for resid {terminalresids[i]} on segment {seg}. Did not apply it."
)
continue
elif cap not in capresname: # If it is defined, test if supported
raise RuntimeError(
f"In segment {seg}, the {cap} cap is not supported. Try using {capresname} instead."
)
# Test if cap is already applied
testcap = np.where(segidm & residm & (mol.resname == cap))[0]
if len(testcap) != 0:
logger.warning(
f"Cap {cap} already exists on segment {seg}. Did not re-apply it."
)
continue
# Test if the atom to change exists
termatomsids = np.zeros(residm.shape, dtype=bool)
for atm in terminalatoms[cap]:
termatomsids |= mol.name == atm
termatomsids = np.where(termatomsids & segidm & residm)[0]
if len(termatomsids) == 0:
# Create new atom
termcaid = np.where(segidm & residm & (mol.name == "CA"))[0]
termcenterid = np.where(segidm & residm
& (mol.name == capatomtype[1 - i]))[0]
atom = Molecule()
atom.empty(1)
atom.record = np.array(["ATOM"],
dtype=Molecule._dtypes["record"])
atom.name = np.array([capatomtype[i]],
dtype=Molecule._dtypes["name"])
atom.resid = np.array([terminalresids[i] - 1 + 2 * i],
dtype=Molecule._dtypes["resid"])
atom.resname = np.array([cap],
dtype=Molecule._dtypes["resname"])
atom.segid = np.array([seg], dtype=Molecule._dtypes["segid"])
atom.element = np.array([capatomtype[i]],
dtype=Molecule._dtypes["element"])
atom.chain = np.array(
[np.unique(mol.chain[segidm])],
dtype=Molecule._dtypes["chain"]
) # TODO: Assumption of single chain in a segment might be wrong
atom.coords = mol.coords[termcenterid] + 0.33 * np.subtract(
mol.coords[termcenterid], mol.coords[termcaid])
mol.insert(atom, terminalids[i])
else:
# Select atom to change, do changes to cap, and change resid
newatom = np.max(termatomsids)
mol.set("resname", cap, sel=newatom)
mol.set("name", capatomtype[i], sel=newatom)
mol.set("element", capatomtype[i], sel=newatom)
mol.set("resid", terminalresids[i] - 1 + 2 * i,
sel=newatom) # if i=0 => resid-1; i=1 => resid+1
# Reorder
neworder = np.arange(mol.numAtoms)
neworder[newatom] = terminalids[i]
neworder[terminalids[i]] = newatom
_reorderMol(mol, neworder)
# For each cap
for i, cap in enumerate(caps[seg]):
if cap is None or (isinstance(cap, str) and cap == "none"):
continue
# Remove lingering hydrogens or oxygens in terminals
mol.remove(
'segid {} and resid "{}" and name {}'.format(
seg, orig_terminalresids[i], " ".join(terminalatoms[cap])),
_logger=False,
)
# Remove terminal hydrogens regardless of caps or no caps. tleap confuses itself when it makes residues into terminals.
# For example HID has an atom H which in NHID should become H[123] and crashes tleap.
for seg in np.unique(mol.get("segid", "protein")):
segidm = mol.segid == seg # Mask for segid
segididx = np.where(segidm)[0]
resids = mol.resid[segididx]
mol.remove(
f'(resid "{resids[0]}" "{resids[-1]}") and segid {seg} and hydrogen',
_logger=False,
)