Source Code for Module Chem.Pharm2D.Matcher

  1  # $Id: Matcher.py 2 2006-05-06 22:54:39Z glandrum $ 
  2  # 
  3  # Copyright (C) 2003-2006 greg Landrum and Rational Discovery LLC 
  4  # 
  5  #   @@ All Rights Reserved  @@ 
  6  # 
  7  """ functionality for finding pharmacophore matches in molecules 
  8   
  9   
 10    See Docs/Chem/Pharm2D.triangles.jpg for an illustration of the way 
 11    pharmacophores are broken into triangles and labelled. 
 12   
 13    See Docs/Chem/Pharm2D.signatures.jpg for an illustration of bit 
 14    numbering 
 15   
 16  """ 
 17  import Chem 
 18  from Chem.Pharm2D import Utils 
 19   
 20  import types 
 21  import exceptions 
 22 -class MatchError(exceptions.Exception): 
 23    pass 
 24   
 25  _verbose = 0 
 26 -def GetAtomsMatchingBit(sig,bitIdx,mol,dMat=None,justOne=0,matchingAtoms=None): 
 27    """ Returns a list of lists of atom indices for a bit 
 28   
 29      **Arguments** 
 30   
 31        - sig: a signature 
 32   
 33        - bitIdx: the bit to be queried 
 34   
 35        - mol: the molecule to be examined 
 36   
 37        - dMat: (optional) the distance matrix of the molecule 
 38   
 39        - justOne: (optional) if this is nonzero, only the first match 
 40          will be returned. 
 41   
 42        - matchingAtoms: (optional) if this is nonzero, it should 
 43          contain a sequence of sequences with the indices of atoms in 
 44          the molecule which match each of the patterns used by the 
 45          signature. 
 46   
 47      **Returns** 
 48   
 49        a list of tuples with the matching atoms 
 50    """ 
 51    assert sig.GetShortestPathsOnly(),'not implemented for non-shortest path signatures' 
 52    nPts,pattCombo,scaffold = sig.GetBitInfo(bitIdx) 
 53    if _verbose: 
 54      print 'info:',nPts 
 55      print '\t',pattCombo 
 56      print '\t',scaffold 
 57     
 58    # find the atoms that match each pattern 
 59    choices = [] 
 60    for pattIdx in pattCombo: 
 61      tmp = None 
 62      if matchingAtoms is not None: 
 63        tmp = matchingAtoms[pattIdx] 
 64      else: 
 65        patt = sig.GetPattern(pattIdx) 
 66        tmp = mol.GetSubstructMatches(patt) 
 67      if tmp: 
 68        choices.append(tmp) 
 69      else: 
 70        # one of the patterns didn't find a match, we 
 71        #  can return now 
 72        if _verbose: print 'no match found for pattern:',pattIdx 
 73        return [] 
 74    
 75    if _verbose: 
 76      print 'choices:' 
 77      print choices 
 78   
 79    if dMat is None: 
 80      useBO = sig.GetIncludeBondOrder() 
 81      dMat = Chem.GetDistanceMatrix(mol,useBO) 
 82   
 83    matches = [] 
 84    distsToCheck = Utils.nPointDistDict[nPts] 
 85   
 86    protoPharmacophores = Utils.GetAllCombinations(choices,noDups=1) 
 87   
 88    res = [] 
 89    for protoPharm in protoPharmacophores: 
 90      if _verbose: print 'protoPharm:',protoPharm 
 91      for i in range(len(distsToCheck)): 
 92        dLow,dHigh = sig.GetBin(scaffold[i]) 
 93        a1,a2 = distsToCheck[i] 
 94        # 
 95        # FIX: this is making all kinds of assumptions about 
 96        #  things being single-atom matches (or at least that 
 97        #  only the first atom matters 
 98        # 
 99        idx1,idx2 = protoPharm[a1][0],protoPharm[a2][0] 
100        dist = dMat[idx1,idx2] 
101        if _verbose: print '\t dist: %d->%d = %d (%d,%d)'%(idx1,idx2,dist,dLow,dHigh) 
102        if dist < dLow or dist >= dHigh: 
103          break 
104      else: 
105        if _verbose: print 'Found one' 
106        # we found it 
107        protoPharm.sort() 
108        protoPharm = tuple(protoPharm) 
109        if protoPharm not in res: 
110          res.append(protoPharm) 
111          if justOne: break 
112    return res 
113   
114  if __name__ == '__main__': 
115    import Chem 
116    from Chem.Pharm2D import SigFactory,Generate 
117         
118    factory = SigFactory.SigFactory() 
119    factory.SetBins([(1,2),(2,5),(5,8)]) 
120    factory.SetPatternsFromSmarts(['O','N']) 
121    factory.SetMinCount(2) 
122    factory.SetMaxCount(3) 
123    sig = factory.GetSignature() 
124     
125    mol = Chem.MolFromSmiles('OCC(=O)CCCN') 
126    Generate.Gen2DFingerprint(mol,sig) 
127    print 'onbits:',list(sig.GetOnBits()) 
128   
129    _verbose=0 
130    for bit in sig.GetOnBits(): 
131      as = GetAtomsMatchingBit(sig,bit,mol) 
132      print '\tBit %d: '%(bit),as 
133   
134       
135    print '--------------------------' 
136    sig = factory.GetSignature() 
137    sig.SetIncludeBondOrder(1) 
138    Generate.Gen2DFingerprint(mol,sig) 
139    print 'onbits:',list(sig.GetOnBits()) 
140   
141    for bit in sig.GetOnBits(): 
142      as = GetAtomsMatchingBit(sig,bit,mol) 
143      print '\tBit %d: '%(bit),as 
144