Python rdkit.Chem.AllChem.MolFromSmarts() Examples

def loadsmarts(fname):
    ret = []
    with open(fname) as f:
        for line in f:
            if line.startswith('#') or len(line.strip()) == 0:
                continue
            tokens = line.split()
            mol = Chem.MolFromSmarts(tokens[0])
            ret.append(mol)
    return ret 

def getDihedralMatches(mol):
    '''return list of atom indices of dihedrals'''
    #this is rdkit's "strict" pattern
    pattern = r"*~[!$(*#*)&!D1&!$(C(F)(F)F)&!$(C(Cl)(Cl)Cl)&!$(C(Br)(Br)Br)&!$(C([CH3])([CH3])[CH3])&!$([CD3](=[N,O,S])-!@[#7,O,S!D1])&!$([#7,O,S!D1]-!@[CD3]=[N,O,S])&!$([CD3](=[N+])-!@[#7!D1])&!$([#7!D1]-!@[CD3]=[N+])]-!@[!$(*#*)&!D1&!$(C(F)(F)F)&!$(C(Cl)(Cl)Cl)&!$(C(Br)(Br)Br)&!$(C([CH3])([CH3])[CH3])]~*"
    qmol = Chem.MolFromSmarts(pattern)
    matches = mol.GetSubstructMatches(qmol);
    #these are all sets of 4 atoms, uniquify by middle two
    uniqmatches = []
    seen = set()
    for (a,b,c,d) in matches:
        if (b,c) not in seen:
            seen.add((b,c))
            uniqmatches.append((a,b,c,d))
    return uniqmatches 

def get_special_groups(mol):
    '''Given an RDKit molecule, this function returns a list of tuples, where
    each tuple contains the AtomIdx's for a special group of atoms which should 
    be included in a fragment all together. This should only be done for the 
    reactants, otherwise the products might end up with mapping mismatches

    We draw a distinction between atoms in groups that trigger that whole
    group to be included, and "unimportant" atoms in the groups that will not
    be included if another atom matches.'''

    # Define templates
    group_templates = [ 
        (range(3), '[OH0,SH0]=C[O,Cl,I,Br,F]',), # carboxylic acid / halogen
        (range(3), '[OH0,SH0]=CN',), # amide/sulfamide
        (range(4), 'S(O)(O)[Cl]',), # sulfonyl chloride
        (range(3), 'B(O)O',), # boronic acid/ester
        ((0,), '[Si](C)(C)C'), # trialkyl silane
        ((0,), '[Si](OC)(OC)(OC)'), # trialkoxy silane, default to methyl
        (range(3), '[N;H0;$(N-[#6]);D2]-,=[N;D2]-,=[N;D1]',), # azide
        (range(8), 'O=C1N([Br,I,F,Cl])C(=O)CC1',), # NBS brominating agent
        (range(11), 'Cc1ccc(S(=O)(=O)O)cc1'), # Tosyl
        ((7,), 'CC(C)(C)OC(=O)[N]'), # N(boc)
        ((4,), '[CH3][CH0]([CH3])([CH3])O'), # 
        (range(2), '[C,N]=[C,N]',), # alkene/imine
        (range(2), '[C,N]#[C,N]',), # alkyne/nitrile
        ((2,), 'C=C-[*]',), # adj to alkene
        ((2,), 'C#C-[*]',), # adj to alkyne
        ((2,), 'O=C-[*]',), # adj to carbonyl
        ((3,), 'O=C([CH3])-[*]'), # adj to methyl ketone
        ((3,), 'O=C([O,N])-[*]',), # adj to carboxylic acid/amide/ester
        (range(4), 'ClS(Cl)=O',), # thionyl chloride
        (range(2), '[Mg,Li,Zn,Sn][Br,Cl,I,F]',), # grinard/metal (non-disassociated)
        (range(3), 'S(O)(O)',), # SO2 group
        (range(2), 'N~N',), # diazo
        ((1,), '[!#6;R]@[#6;R]',), # adjacency to heteroatom in ring
        ((2,), '[a!c]:a:a',), # two-steps away from heteroatom in aromatic ring
        #((1,), 'c(-,=[*]):c([Cl,I,Br,F])',), # ortho to halogen on ring - too specific?
        #((1,), 'c(-,=[*]):c:c([Cl,I,Br,F])',), # meta to halogen on ring - too specific?
        ((0,), '[B,C](F)(F)F'), # CF3, BF3 should have the F3 included
    ]

    # Stereo-specific ones (where we will need to include neighbors)
    # Tetrahedral centers should already be okay...
    group_templates += [
        ((1,2,), '[*]/[CH]=[CH]/[*]'), # trans with two hydrogens
        ((1,2,), '[*]/[CH]=[CH]\[*]'), # cis with two hydrogens
        ((1,2,), '[*]/[CH]=[CH0]([*])\[*]'), # trans with one hydrogens
        ((1,2,), '[*]/[D3;H1]=[!D1]'), # specified on one end, can be N or C
    ]

    # Build list
    groups = []
    for (add_if_match, template) in group_templates:
        matches = mol.GetSubstructMatches(Chem.MolFromSmarts(template), useChirality=True)
        for match in matches:
            add_if = []
            for pattern_idx, atom_idx in enumerate(match):
                if pattern_idx in add_if_match:
                    add_if.append(atom_idx)
            groups.append((add_if, match))
    return groups 

def get_special_groups(mol):
    '''Given an RDKit molecule, this function returns a list of tuples, where
    each tuple contains the AtomIdx's for a special group of atoms which should 
    be included in a fragment all together. This should only be done for the 
    reactants, otherwise the products might end up with mapping mismatches

    We draw a distinction between atoms in groups that trigger that whole
    group to be included, and "unimportant" atoms in the groups that will not
    be included if another atom matches.'''

    # Define templates, based on Functional_Group_Hierarchy.txt from Greg Laandrum
    group_templates = [ 
        (range(3), '[OH0,SH0]=C[O,Cl,I,Br,F]',), # carboxylic acid / halogen
        (range(3), '[OH0,SH0]=CN',), # amide/sulfamide
        (range(4), 'S(O)(O)[Cl]',), # sulfonyl chloride
        (range(3), 'B(O)O',), # boronic acid/ester
        ((0,), '[Si](C)(C)C'), # trialkyl silane
        ((0,), '[Si](OC)(OC)(OC)'), # trialkoxy silane, default to methyl
        (range(3), '[N;H0;$(N-[#6]);D2]-,=[N;D2]-,=[N;D1]',), # azide
        (range(8), 'O=C1N([Br,I,F,Cl])C(=O)CC1',), # NBS brominating agent
        (range(11), 'Cc1ccc(S(=O)(=O)O)cc1'), # Tosyl
        ((7,), 'CC(C)(C)OC(=O)[N]'), # N(boc)
        ((4,), '[CH3][CH0]([CH3])([CH3])O'), # 
        (range(2), '[C,N]=[C,N]',), # alkene/imine
        (range(2), '[C,N]#[C,N]',), # alkyne/nitrile
        ((2,), 'C=C-[*]',), # adj to alkene
        ((2,), 'C#C-[*]',), # adj to alkyne
        ((2,), 'O=C-[*]',), # adj to carbonyl
        ((3,), 'O=C([CH3])-[*]'), # adj to methyl ketone
        ((3,), 'O=C([O,N])-[*]',), # adj to carboxylic acid/amide/ester
        (range(4), 'ClS(Cl)=O',), # thionyl chloride
        (range(2), '[Mg,Li,Zn,Sn][Br,Cl,I,F]',), # grinard/metal (non-disassociated)
        (range(3), 'S(O)(O)',), # SO2 group
        (range(2), 'N~N',), # diazo
        ((1,), '[!#6;R]@[#6;R]',), # adjacency to heteroatom in ring
        ((2,), '[a!c]:a:a',), # two-steps away from heteroatom in aromatic ring
        #((1,), 'c(-,=[*]):c([Cl,I,Br,F])',), # ortho to halogen on ring - too specific?
        #((1,), 'c(-,=[*]):c:c([Cl,I,Br,F])',), # meta to halogen on ring - too specific?
        ((0,), '[B,C](F)(F)F'), # CF3, BF3 should have the F3 included
    ]

    # Stereo-specific ones (where we will need to include neighbors)
    # Tetrahedral centers should already be okay...
    group_templates += [
        ((1,2,), '[*]/[CH]=[CH]/[*]'), # trans with two hydrogens
        ((1,2,), '[*]/[CH]=[CH]\[*]'), # cis with two hydrogens
        ((1,2,), '[*]/[CH]=[CH0]([*])\[*]'), # trans with one hydrogens
        ((1,2,), '[*]/[D3;H1]=[!D1]'), # specified on one end, can be N or C
    ]

    # Build list
    groups = []
    for (add_if_match, template) in group_templates:
        matches = mol.GetSubstructMatches(Chem.MolFromSmarts(template), useChirality=True)
        for match in matches:
            add_if = []
            for pattern_idx, atom_idx in enumerate(match):
                if pattern_idx in add_if_match:
                    add_if.append(atom_idx)
            groups.append((add_if, match))
    return groups