def alignSequencesByChain(PDBs, **kwargs):
"""
Runs buildMSA for each chain and optionally joins the results.
Returns either a single MSA or a dictionary containing an MSA for each chain.
:arg PDBs: a list or array of :class:`AtomGroup` objects or PDB IDs
a mixed list containing both is acceptable
:type PDBs: list or :class:`~numpy.ndarray`
:arg join_chains: whether to join chain alignments
default is True
:type join_chains: bool
:arg join_char: a character for joining chain alignments
default is '/' as used by PIR format alignments
:type join_char: str
"""
if not (isinstance(PDBs, list) or isinstance(PDBs, ndarray)):
raise TypeError('PDBs should be a list or array')
if PDBs == []:
raise ValueError('PDBs should not be an empty list')
pdbs = []
chains = []
for i, pdb in enumerate(PDBs):
if isinstance(pdb, Atomic):
pdbs.append(pdb)
else:
raise TypeError(
'each entry in PDBs must be a :class:`Atomic` instance')
chains.append([])
for chain in list(pdbs[i].getHierView()):
chains[i].append(chain)
if i != 0 and len(chains[i]) != len(chains[0]):
raise ValueError('all pdbs should have the same number of chains')
labels = []
for pdb in pdbs:
chids = ''
for chain in list(pdb.getHierView()):
chids += chain.getChid()
labels.append(pdb.getTitle().split('_')[0] + '_' + chids)
chains = array(chains)
chain_alignments = []
alignments = {}
labels_lists = []
for j in range(len(chains[0])):
prefix = 'chain_' + chains[0, j].getChid()
msa = buildMSA(chains[:, j], title=prefix, labels=labels)
# make all alignments have the sequences in the same order as the 0th
labels_lists.append([])
for sequence in msa:
labels_lists[j].append(sequence.getLabel())
if j > 0:
msaarr = []
for label in labels_lists[0]:
msaarr.append(msa.getArray()[msa.getIndex(label)])
msaarr = array(msaarr)
msa = MSA(msaarr,
title='reordered_msa_1',
labels=list(labels_lists[0]))
writeMSA(prefix + '.aln', msa)
chain_alignments.append(msa)
# after reordering, create the alignments dictionary
alignments[labels_lists[0][0].split('_')[1][j]] = msa
join_chains = kwargs.get('join_chains', True)
join_char = kwargs.get('join_char', '/')
if join_chains:
aligned_sequences = list(zeros(shape(chain_alignments)).T)
for j in range(shape(chain_alignments)[1]):
aligned_sequences[j] = list(aligned_sequences[j])
orig_labels = []
for i, chain_alignment in enumerate(chain_alignments):
for j, sequence in enumerate(chain_alignment):
aligned_sequences[j][i] = str(sequence)
if i == 0:
orig_labels.append(sequence.getLabel())
joined_msaarr = []
for j in range(shape(chain_alignments)[1]):
joined_msaarr.append(
array(list(join_char.join(aligned_sequences[j]))))
joined_msaarr = array(joined_msaarr)
result = MSA(joined_msaarr, title='joined_chains', labels=orig_labels)
result = refineMSA(result, colocc=1e-9) # remove gap-only cols
else:
result = alignments
return result
def alignSequencesByChain(PDBs, **kwargs):
"""
Runs :func:`buildMSA` for each chain and optionally joins the results.
Returns either a single :class:`MSA` or a dictionary containing an :class:`MSA` for each chain.
:arg PDBs: a list of :class:`AtomGroup` objects
:type PDBs: list
:arg join_chains: whether to join chain alignments
default is True
:type join_chains: bool
:arg join_char: a character for joining chain alignments
default is '/' as used by PIR format alignments
:type join_char: str
"""
if isscalar(PDBs):
raise TypeError('PDBs should be array-like')
if not PDBs:
raise ValueError('PDBs should not be empty')
pdbs = []
chains = []
for i, pdb in enumerate(PDBs):
if isinstance(pdb, Atomic):
pdbs.append(pdb)
else:
raise TypeError(
'each entry in PDBs must be a :class:`Atomic` instance')
chains.append([])
for chain in list(pdbs[i].getHierView()):
chains[i].append(chain)
if i != 0 and len(chains[i]) != len(chains[0]):
raise ValueError('all pdbs should have the same number of chains')
labels = []
for pdb in pdbs:
chids = ''
for chain in list(pdb.getHierView()):
chids += chain.getChid()
labels.append(pdb.getTitle() + '_' + chids)
chains = array(chains)
chain_alignments = []
alignments = {}
for j in range(len(chains[0])):
prefix = 'chain_' + chains[0, j].getChid()
msa = buildMSA(chains[:, j], title=prefix, labels=labels)
msa = refineMSA(msa, colocc=1e-9) # remove gap-only cols
chain_alignments.append(msa)
alignments[labels[0].split('_')[1][j]] = msa
join_chains = kwargs.get('join_chains', True)
join_char = kwargs.get('join_char', '/')
if len(chains[0]) == 1:
join_chains = False
if join_chains:
joined_msaarr = []
for i, chain_alignment in enumerate(chain_alignments):
pdb_seqs = []
for j, sequence in enumerate(chain_alignment):
pdb_seqs.append(sequence)
joined_msaarr.append(join_char.join(pdb_seqs))
result = MSA(joined_msaarr,
title='joined_chains',
labels=[label.split('_')[0] for label in labels])
else:
result = alignments
if len(result) == 1:
result = result[list(result.keys())[0]]
return result
def alignSequencesByChain(PDBs, **kwargs):
"""
Runs :func:`buildMSA` for each chain and optionally joins the results.
Returns either a single :class:`MSA` or a dictionary containing an :class:`MSA` for each chain.
:arg PDBs: a list of :class:`AtomGroup` objects
:type PDBs: list
:arg join_chains: whether to join chain alignments
default is True
:type join_chains: bool
:arg join_char: a character for joining chain alignments
default is '/' as used by PIR format alignments
:type join_char: str
"""
if isscalar(PDBs):
raise TypeError('PDBs should be array-like')
if not PDBs:
raise ValueError('PDBs should not be empty')
pdbs = []
chains = []
for i, pdb in enumerate(PDBs):
if isinstance(pdb, Atomic):
pdbs.append(pdb)
else:
raise TypeError('each entry in PDBs must be a :class:`Atomic` instance')
chains.append([])
for chain in list(pdbs[i].getHierView()):
chains[i].append(chain)
if i != 0 and len(chains[i]) != len(chains[0]):
raise ValueError('all pdbs should have the same number of chains')
labels = []
for pdb in pdbs:
chids = ''
for chain in list(pdb.getHierView()):
chids += chain.getChid()
labels.append(pdb.getTitle() + '_' + chids)
chains = array(chains)
chain_alignments = []
alignments = {}
for j in range(len(chains[0])):
prefix = 'chain_' + chains[0, j].getChid()
msa = buildMSA(chains[:, j], title=prefix, labels=labels)
msa = refineMSA(msa, colocc=1e-9) # remove gap-only cols
chain_alignments.append(msa)
alignments[labels[0].split('_')[1][j]] = msa
join_chains = kwargs.get('join_chains', True)
join_char = kwargs.get('join_char', '/')
if len(chains[0]) == 1:
join_chains = False
if join_chains:
joined_msaarr = []
for i, chain_alignment in enumerate(chain_alignments):
pdb_seqs = []
for j, sequence in enumerate(chain_alignment):
pdb_seqs.append(sequence)
joined_msaarr.append(join_char.join(pdb_seqs))
result = MSA(joined_msaarr, title='joined_chains',
labels=[label.split('_')[0] for label in labels])
else:
result = alignments
if len(result) == 1:
result = result[list(result.keys())[0]]
return result
