def evol_occupancy(msa, **kwargs):
from numpy import arange
import prody
from prody import parseMSA, calcMSAOccupancy, showMSAOccupancy, writeArray
from os.path import splitext
prefix = kwargs.get('prefix')
if prefix is None:
prefix, _ = splitext(msa)
if _.lower() == '.gz':
prefix, _ = splitext(prefix)
prefix += '_occupancy'
msa = parseMSA(msa)
numformat = kwargs.get('numformat', '%12g')
occupancy, suffix = [], []
occaxis = kwargs.get('occaxis', 'row')
if occaxis == 'both':
suffix = ['_row', '_col']
occupancy.append(calcMSAOccupancy(msa, occ='row'))
occupancy.append(calcMSAOccupancy(msa, occ='col'))
else:
suffix = '_' + occaxis
occupancy.append(calcMSAOccupancy(msa, occ=occaxis))
for i, occ in enumerate(occupancy):
writeArray((prefix + suffix[i] + '.txt'), occ, format=numformat)
for i, occ in enumerate(occupancy):
if kwargs.get('figocc'):
try:
import matplotlib.pyplot as plt
except ImportError:
LOGGER.warn('Matplotlib could not be imported, '
'figures are not saved.')
else:
prody.SETTINGS['auto_show'] = False
width = kwargs.get('figwidth', 8)
height = kwargs.get('figheight', 6)
xlabel = kwargs.get('xlabel')
title = kwargs.get('title')
figure = plt.figure(figsize=(width, height))
label = kwargs.get('label')
show = showMSAOccupancy(msa=msa,
occ=occ,
label=label,
xlabel=xlabel,
title=title)
format = kwargs.get('figformat', 'pdf')
figure.savefig(prefix + suffix[i] + '.' + format,
format=format,
dpi=kwargs.get('figdpi', 300))
def evol_occupancy(msa, **kwargs):
from numpy import arange
import prody
from prody import parseMSA, calcMSAOccupancy, showMSAOccupancy, writeArray
from os.path import splitext
prefix = kwargs.get('prefix')
if prefix is None:
prefix, _ = splitext(msa)
if _.lower() == '.gz':
prefix, _ = splitext(prefix)
prefix += '_occupancy'
msa = parseMSA(msa)
numformat = kwargs.get('numformat', '%12g')
occupancy , suffix = [], []
occaxis = kwargs.get('occaxis', 'row')
if occaxis == 'both':
suffix = ['_row', '_col']
occupancy.append(calcMSAOccupancy(msa, occ='row'))
occupancy.append(calcMSAOccupancy(msa, occ='col'))
else:
suffix = '_' + occaxis
occupancy.append(calcMSAOccupancy(msa, occ=occaxis))
for i, occ in enumerate(occupancy):
writeArray((prefix + suffix[i] + '.txt'), occ, format=numformat)
for i, occ in enumerate(occupancy):
if kwargs.get('figocc'):
try:
import matplotlib.pyplot as plt
except ImportError:
LOGGER.warn('Matplotlib could not be imported, '
'figures are not saved.')
else:
prody.SETTINGS['auto_show'] = False
width = kwargs.get('figwidth', 8)
height = kwargs.get('figheight', 6)
xlabel = kwargs.get('xlabel')
title = kwargs.get('title')
figure = plt.figure(figsize=(width, height))
label = kwargs.get('label')
show = showMSAOccupancy(msa=msa, occ=occ, label=label,
xlabel=xlabel, title=title)
format = kwargs.get('figformat', 'pdf')
figure.savefig(prefix + suffix[i] + '.' + format, format=format,
dpi=kwargs.get('figdpi', 300))
def testAll(self):
rowocc = 0.9
colocc = 0.9
seqid = 0.98
label = "FSHB_BOVIN"
refined = refineMSA(FASTA, label=label, seqid=seqid, rowocc=rowocc, colocc=colocc)
index = FASTA.getIndex(label)
which = FASTA_ALPHA[index].nonzero()[0]
expected = FASTA._getArray().take(which, 1)
expected = expected[uniqueSequences(expected, seqid)]
expected = expected[calcMSAOccupancy(expected, "row") >= rowocc]
which = (calcMSAOccupancy(expected) >= colocc).nonzero()[0]
expected = expected.take(which, 1)
assert_array_equal(refined._getArray(), expected)
def testAll(self):
rowocc = 0.9
colocc = 0.9
seqid = 0.98
label = 'FSHB_BOVIN'
refined = refineMSA(FASTA, label=label, seqid=seqid,
rowocc=rowocc, colocc=colocc)
index = FASTA.getIndex(label)
which = FASTA_ALPHA[index].nonzero()[0]
expected = FASTA._getArray().take(which, 1)
expected = expected[uniqueSequences(expected, seqid)]
expected = expected[calcMSAOccupancy(expected, 'row') >= rowocc]
which = (calcMSAOccupancy(expected) >= colocc).nonzero()[0]
expected = expected.take(which, 1)
assert_array_equal(refined._getArray(), expected)
def evol_rankorder(mutinfo, **kwargs):
from prody import parseMSA, LOGGER, parsePDB, calcMSAOccupancy
from prody.utilities import openFile
from os.path import splitext
delimiter = kwargs.get('delimiter')
mi = np.loadtxt(str(mutinfo), delimiter=delimiter)
ndim, shape = mi.ndim, mi.shape
if ndim != 2 or shape[0] != shape[1]:
raise ValueError('mutinfo must contain a square matrix')
msa, label = kwargs.get('msa'), kwargs.get('label')
pdb, pdbflag = kwargs.get('pdb'), False
resnum = None
if pdb is not None:
from prody import parsePDB
try:
pdb = parsePDB(pdb)
except:
LOGGER.info('Could not parse PDB, ignoring PDB input')
else:
chains = list(pdb.iterChains())
for chain in chains:
sel = chain.select('protein and name CA')
if sel.numAtoms() == shape[0]:
resnum = sel.getResnums()
coordset = sel.getCoordsets()
distance = calcAllDist(coordset)
pdbflag = True
label = pdb.getTitle()
LOGGER.info('Residue numbers will be based on pdb: '
'{0}'.format(pdb.getTitle()))
break
else:
LOGGER.info('Number of residues in PDB does not match '
'mutinfo matrix, ignoring PDB input')
if not pdbflag:
if msa is not None:
msa = parseMSA(msa)
if msa.numResidues() != shape[0]:
LOGGER.info('Input MSA and mutinfo do not have similar no '
'of residues, ignoring MSA')
else:
index = msa.getIndex(label)
if index is None:
if label is not None:
LOGGER.info('Could not find given label in MSA, '
'using complete sequence from MSA')
occ = calcMSAOccupancy(msa._msa, 'row')
index = np.where(occ == occ.max())[0][0]
label, seq, start, end = msa[index]
else:
label, seq, start, end = msa[index]
if (start and end is not None) and (start < end):
resnum = np.arange(start, end+1)
if len(resnum) != shape[0]:
LOGGER.info('Label: {0}/{1}-{2} and mutinfo do '
'not have similar no of residues, using '
'serial indexing'.format(label, start, end))
label = 'Serial Index'
resnum = np.arange(1, shape[0]+1)
else:
LOGGER.info('Residue numbers will be based on label: '
'{0}'.format(label))
else:
LOGGER.info('Could not identify residue indexes from MSA'
' using serial indexing')
label = 'Serial Index'
resnum = np.arange(1, shape[0]+1)
else:
LOGGER.info('MSA or PDB not given or does not match mutinfo, '
'using serial indexing')
resnum = np.arange(1, shape[0]+1)
LOGGER.info('Residue numbers start and end with {0}-{1}'.
format(str(resnum[0]), str(resnum[-1])))
outname = kwargs.get('outname')
if outname is None:
outname, ext = splitext(str(mutinfo))
if ext.lower() == '.gz':
outname, _ = splitext(str(mutinfo))
else:
outname, ext = splitext(str(outname))
if ext is None:
ext = '.txt'
outname += '_rankorder' + ext
zscore = kwargs.get('zscore')
if zscore:
LOGGER.info('zscore normalization applied such that each column '
'has 0 mean and standard deviation 1')
header = 'Serial\tRow\tColumn\tZscore'
mi = (mi - mi.mean(0)) / mi.std(0)
else:
header = 'Serial\tRow\tColumn\tMI'
mi_ind_start, mi_ind_end = np.tril_indices(shape[0], k=-1)
mi_matrix = mi[mi_ind_start, mi_ind_end]
sorted_index = mi_matrix.argsort(axis=None)[::-1]
row = mi_ind_start[sorted_index]
column = mi_ind_end[sorted_index]
count = 1
i = 0
f = openFile(outname, 'wb')
if label is None:
label = 'Serial Index'
numpairs = kwargs.get('numpairs')
size = len(row)
seqsep = kwargs.get('seqsep')
if not kwargs.get('usedist') or not pdbflag:
if kwargs.get('usedist'):
LOGGER.info('use-struct-sep set to true, but PDB not given or '
'incorrect residue number. Using sequence separation')
else:
if pdbflag:
LOGGER.info('use-dist not set, using sequence separation'
' to report coevolving pairs')
f.write(('Label: '+ label + '\t' + 'Residue Numbers: ' +
str(resnum[0]) + '-' + str(resnum[-1]) + '\tSequence Separation:' +
str(seqsep) + '\n'))
if pdbflag:
f.write((header + '\tDistance\n'))
while count <=numpairs and i < size:
if row[i] > (column[i] + seqsep):
f.write('{0}\t{1}\t{2}\t{3:.3f}\t{4:.2f}\n'.
format(count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]],
distance[row[i], column[i]]))
count += 1
i += 1
else:
f.write((header + '\n'))
while count <=numpairs and i < size:
if row[i] > (column[i] + seqsep):
f.write('{0}\t{1}\t{2}\t{3:.3f}\n'.
format(count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]]))
count += 1
i += 1
else:
structsep = kwargs.get('dist')
f.write(('Label: '+ label + '\t' + 'Residue Numbers: ' +
str(resnum[0]) + '-' + str(resnum[-1]) + 'Distance Cutoff:' +
str(structsep) + '\n'))
f.write((header + '\tDistance\n'))
while count <=numpairs and i < size:
if distance[row[i], column[i]] > structsep:
f.write('{0}\t{1}\t{2}\t{3:.3f}\t{4:.2f}\n'.
format(count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]],
distance[row[i], column[i]]))
count += 1
i += 1
f.close()
def testSequenceOccupancy(self):
assert_array_equal(calcMSAOccupancy(FASTA, 'sequence'),
FASTA_ALPHA.sum(1) / (FASTA.numResidues() * 1.0))
def testResidueOccupancy(self):
assert_array_equal(calcMSAOccupancy(FASTA, 'residue'),
FASTA_ALPHA.sum(0) / (FASTA.numSequences() * 1.0))
def testSequenceCount(self):
assert_array_equal(calcMSAOccupancy(FASTA, 'sequence', count=1),
FASTA_ALPHA.sum(1))
def testResidueCount(self):
assert_array_equal(calcMSAOccupancy(FASTA, 'residue', count=1),
FASTA_ALPHA.sum(0))
def testSequenceOccupancy(self):
assert_array_equal(calcMSAOccupancy(FASTA, "sequence"), FASTA_ALPHA.sum(1) / (FASTA.numResidues() * 1.0))
def testResidueOccupancy(self):
assert_array_equal(calcMSAOccupancy(FASTA, "residue"), FASTA_ALPHA.sum(0) / (FASTA.numSequences() * 1.0))
def testSequenceCount(self):
assert_array_equal(calcMSAOccupancy(FASTA, "sequence", count=1), FASTA_ALPHA.sum(1))
def testResidueCount(self):
assert_array_equal(calcMSAOccupancy(FASTA, "residue", count=1), FASTA_ALPHA.sum(0))
def evol_rankorder(mutinfo, **kwargs):
from prody import parseMSA, LOGGER, parsePDB, calcMSAOccupancy
from prody.utilities import openFile
from os.path import splitext
delimiter = kwargs.get('delimiter')
mi = np.loadtxt(str(mutinfo), delimiter=delimiter)
ndim, shape = mi.ndim, mi.shape
if ndim != 2 or shape[0] != shape[1]:
raise ValueError('mutinfo must contain a square matrix')
msa, label = kwargs.get('msa'), kwargs.get('label')
pdb, pdbflag = kwargs.get('pdb'), False
resnum = None
if pdb is not None:
from prody import parsePDB
try:
pdb = parsePDB(pdb)
except:
LOGGER.info('Could not parse PDB, ignoring PDB input')
else:
chains = list(pdb.iterChains())
for chain in chains:
sel = chain.select('protein and name CA')
if sel.numAtoms() == shape[0]:
resnum = sel.getResnums()
coordset = sel.getCoordsets()
distance = calcAllDist(coordset)
pdbflag = True
label = pdb.getTitle()
LOGGER.info('Residue numbers will be based on pdb: '
'{0}'.format(pdb.getTitle()))
break
else:
LOGGER.info('Number of residues in PDB does not match '
'mutinfo matrix, ignoring PDB input')
if not pdbflag:
if msa is not None:
msa = parseMSA(msa)
if msa.numResidues() != shape[0]:
LOGGER.info('Input MSA and mutinfo do not have similar no '
'of residues, ignoring MSA')
else:
index = msa.getIndex(label)
if index is None:
if label is not None:
LOGGER.info('Could not find given label in MSA, '
'using complete sequence from MSA')
occ = calcMSAOccupancy(msa._msa, 'row')
index = np.where(occ == occ.max())[0][0]
label, seq, start, end = msa[index]
else:
label, seq, start, end = msa[index]
if (start and end is not None) and (start < end):
resnum = np.arange(start, end + 1)
if len(resnum) != shape[0]:
LOGGER.info('Label: {0}/{1}-{2} and mutinfo do '
'not have similar no of residues, using '
'serial indexing'.format(
label, start, end))
label = 'Serial Index'
resnum = np.arange(1, shape[0] + 1)
else:
LOGGER.info('Residue numbers will be based on label: '
'{0}'.format(label))
else:
LOGGER.info('Could not identify residue indexes from MSA'
' using serial indexing')
label = 'Serial Index'
resnum = np.arange(1, shape[0] + 1)
else:
LOGGER.info('MSA or PDB not given or does not match mutinfo, '
'using serial indexing')
resnum = np.arange(1, shape[0] + 1)
LOGGER.info('Residue numbers start and end with {0}-{1}'.format(
str(resnum[0]), str(resnum[-1])))
outname = kwargs.get('outname')
if outname is None:
outname, ext = splitext(str(mutinfo))
if ext.lower() == '.gz':
outname, _ = splitext(str(mutinfo))
else:
outname, ext = splitext(str(outname))
if ext is None:
ext = '.txt'
outname += '_rankorder' + ext
zscore = kwargs.get('zscore')
if zscore:
LOGGER.info('zscore normalization applied such that each column '
'has 0 mean and standard deviation 1')
header = 'Serial\tRow\tColumn\tZscore'
mi = (mi - mi.mean(0)) / mi.std(0)
else:
header = 'Serial\tRow\tColumn\tMI'
mi_ind_start, mi_ind_end = np.tril_indices(shape[0], k=-1)
mi_matrix = mi[mi_ind_start, mi_ind_end]
sorted_index = mi_matrix.argsort(axis=None)[::-1]
row = mi_ind_start[sorted_index]
column = mi_ind_end[sorted_index]
count = 1
i = 0
f = openFile(outname, 'wb')
if label is None:
label = 'Serial Index'
numpairs = kwargs.get('numpairs')
size = len(row)
seqsep = kwargs.get('seqsep')
if not kwargs.get('usedist') or not pdbflag:
if kwargs.get('usedist'):
LOGGER.info('use-struct-sep set to true, but PDB not given or '
'incorrect residue number. Using sequence separation')
else:
if pdbflag:
LOGGER.info('use-dist not set, using sequence separation'
' to report coevolving pairs')
f.write(('Label: ' + label + '\t' + 'Residue Numbers: ' +
str(resnum[0]) + '-' + str(resnum[-1]) +
'\tSequence Separation:' + str(seqsep) + '\n'))
if pdbflag:
f.write((header + '\tDistance\n'))
while count <= numpairs and i < size:
if row[i] > (column[i] + seqsep):
f.write('{0}\t{1}\t{2}\t{3:.3f}\t{4:.2f}\n'.format(
count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]], distance[row[i], column[i]]))
count += 1
i += 1
else:
f.write((header + '\n'))
while count <= numpairs and i < size:
if row[i] > (column[i] + seqsep):
f.write('{0}\t{1}\t{2}\t{3:.3f}\n'.format(
count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]]))
count += 1
i += 1
else:
structsep = kwargs.get('dist')
f.write(('Label: ' + label + '\t' + 'Residue Numbers: ' +
str(resnum[0]) + '-' + str(resnum[-1]) + 'Distance Cutoff:' +
str(structsep) + '\n'))
f.write((header + '\tDistance\n'))
while count <= numpairs and i < size:
if distance[row[i], column[i]] > structsep:
f.write('{0}\t{1}\t{2}\t{3:.3f}\t{4:.2f}\n'.format(
count, resnum[row[i]], resnum[column[i]],
mi[row[i], column[i]], distance[row[i], column[i]]))
count += 1
i += 1
f.close()