def GetBitInfo(self, idx):
""" returns information about the given bit
**Arguments**
- idx: the bit index to be considered
**Returns**
a 3-tuple:
1) the number of points in the pharmacophore
2) the proto-pharmacophore (tuple of pattern indices)
3) the scaffold (tuple of distance indices)
"""
if idx >= self._sigSize:
raise IndexError('bad index (%d) queried. %d is the max' %
(idx, self._sigSize))
# first figure out how many points are in the p'cophore
nPts = self.minPointCount
while nPts < self.maxPointCount and self._starts[nPts + 1] <= idx:
nPts += 1
# how far are we in from the start point?
offsetFromStart = idx - self._starts[nPts]
if _verbose:
print('\t %d Points, %d offset' % (nPts, offsetFromStart))
# lookup the number of scaffolds
nDists = len(Utils.nPointDistDict[nPts])
scaffolds = self._scaffolds[nDists]
nScaffolds = len(scaffolds)
# figure out to which proto-pharmacophore we belong:
protoIdx = offsetFromStart // nScaffolds
indexCombos = Utils.GetIndexCombinations(self._nFeats, nPts)
combo = tuple(indexCombos[protoIdx])
if _verbose:
print('\t combo: %s' % (str(combo)))
# and which scaffold:
scaffoldIdx = offsetFromStart % nScaffolds
scaffold = scaffolds[scaffoldIdx]
if _verbose:
print('\t scaffold: %s' % (str(scaffold)))
return nPts, combo, scaffold
def Gen2DFingerprint(mol, sigFactory, perms=None, dMat=None):
""" generates a 2D fingerprint for a molecule using the
parameters in _sig_
**Arguments**
- mol: the molecule for which the signature should be generated
- sigFactory : the SigFactory object with signature parameters
NOTE: no preprocessing is carried out for _sigFactory_.
It *must* be pre-initialized.
- perms: (optional) a sequence of permutation indices limiting which
pharmacophore combinations are allowed
- dMat: (optional) the distance matrix to be used
"""
if not isinstance(sigFactory, SigFactory.SigFactory):
raise ValueError('bad factory')
featFamilies = sigFactory.GetFeatFamilies()
if _verbose:
print('* feat famillies:', featFamilies)
nFeats = len(featFamilies)
minCount = sigFactory.minPointCount
maxCount = sigFactory.maxPointCount
if maxCount > 3:
logger.warning(
' Pharmacophores with more than 3 points are not currently supported.\nSetting maxCount to 3.'
)
maxCount = 3
# generate the molecule's distance matrix, if required
if dMat is None:
from rdkit import Chem
useBO = sigFactory.includeBondOrder
dMat = Chem.GetDistanceMatrix(mol, useBO)
# generate the permutations, if required
if perms is None:
perms = []
for count in range(minCount, maxCount + 1):
perms += Utils.GetIndexCombinations(nFeats, count)
# generate the matches:
featMatches = sigFactory.GetMolFeats(mol)
if _verbose:
print(' featMatches:', featMatches)
sig = sigFactory.GetSignature()
for perm in perms:
# the permutation is a combination of feature indices
# defining the feature set for a proto-pharmacophore
featClasses = [0]
for i in range(1, len(perm)):
if perm[i] == perm[i - 1]:
featClasses.append(featClasses[-1])
else:
featClasses.append(featClasses[-1] + 1)
# Get a set of matches at each index of
# the proto-pharmacophore.
matchPerms = [featMatches[x] for x in perm]
if _verbose:
print('\n->Perm: %s' % (str(perm)))
print(' matchPerms: %s' % (str(matchPerms)))
# Get all unique combinations of those possible matches:
matchesToMap = Utils.GetUniqueCombinations(matchPerms, featClasses)
for i, entry in enumerate(matchesToMap):
entry = [x[1] for x in entry]
matchesToMap[i] = entry
if _verbose:
print(' mtM:', matchesToMap)
for match in matchesToMap:
if sigFactory.shortestPathsOnly:
_ShortestPathsMatch(match, perm, sig, dMat, sigFactory)
return sig