def get_next_murcko_fragments(murcko_scaffold, break_fused_rings=True):
"""
Fragment a scaffold into its next set of Murcko fragments.
The fragmenter assumes that a Murcko scaffold is supplied. If not
the behaviour may be undesirable.
Parameters
----------
murcko_scaffold : rdkit.Chem.rdchem.Mol
A Murcko scaffold to fragment
break_fused_rings : bool, optional
If True dissect fused rings. The default is True.
Returns
-------
parents : list
A list of parent scaffolds (next hierarchy [num_rings - 1])
"""
rdlogger.setLevel(4)
if break_fused_rings:
fragmenter = MurckoRingFragmenter()
else:
fragmenter = MurckoRingSystemFragmenter()
parents = [
f.mol for f in set(fragmenter.fragment(Scaffold(murcko_scaffold)))
]
rdlogger.setLevel(3)
return parents
def get_next_murcko_fragments(murcko_scaffold, break_fused_rings=True):
"""Fragment a scaffold into its next set of murcko fragments.
The fragmenter assumes that a murcko scaffold is supplied.
Parameters
----------
murcko_scaffold (Chem.Mol): An rdkit Mol containing a murcko scaffold
break_fused_rings (bool): If True dissect fused rings (default: True)
Returns
-------
parents (list): a list of parent scaffolds (next hierarchy [num_rings - 1])
"""
rdlogger.setLevel(4)
if break_fused_rings:
fragmenter = MurckoRingFragmenter()
else:
fragmenter = MurckoRingSystemFragmenter()
parents = [
f.mol for f in set(fragmenter.fragment(Scaffold(murcko_scaffold)))
]
rdlogger.setLevel(3)
return parents
def get_all_murcko_fragments(mol, break_fused_rings=True):
"""
Get all possible murcko fragments from a molecule through
recursive removal of peripheral rings.
Parameters
----------
mol : rdkit.Chem.rdchem.Mol
break_fused_rings : bool, optional
If True dissect fused rings. The default is True.
Returns
-------
list
A list of Murcko fragments for the input molecule.
Examples
--------
Generating Murcko fragments:
>>> from rdkit import Chem
>>> smiles = 'Cc1[nH]cnc1Cn1cccc(-c2ccccc2O)c1=O'
>>> molecule = Chem.MolFromSmiles(smiles)
>>> frags = get_all_murcko_fragments(molecule)
"""
rdlogger.setLevel(4)
if break_fused_rings:
fragmenter = MurckoRingFragmenter()
else:
fragmenter = MurckoRingSystemFragmenter()
mol = get_murcko_scaffold(mol)
rdmolops.RemoveStereochemistry(mol)
scaffold = Scaffold(mol)
parents = {scaffold}
def recursive_generation(child):
for parent in fragmenter.fragment(child):
if parent in parents:
continue
parents.add(parent)
recursive_generation(parent)
recursive_generation(scaffold)
rdlogger.setLevel(3)
return [f.mol for f in parents]
def get_scaffold_frags(frag, hash_func=None):
"""Get fragments from a disconnected structure.
This function is used primarily used during molecular fragmentation.
Parameters
----------
frag : rdkit.Chem.rdchem.Mol
An rdkit Mol containing disconeccted structures.
hash_func : callable, optional
Scaffold hash function, the default is the canonical smiles.
Returns
-------
set
A set of scaffoldgraph.core.Scaffold objects which are unique
fragments obtained from the disconnected structures in the
input molecule.
Notes
-----
The function also performs a `partial sanitization` of the input.
If the molecule fails to sanitize the function will return an
empty set. This case can occur when an aromatic ring system is
dissected in a way in which the resultant system is no longer
aromatic.
"""
try:
# frag.ClearComputedProps()
# frag.UpdatePropertyCache()
# FastFindRings(frag)
partial_sanitization(frag)
except ValueError as e:
# This error is caught as dissecting an aromatic ring system,
# may lead to an undefined state where the resultant system
# is no longer aromatic. We make no attempt to prevent this
# but log it for reference.
# This behaviour may be desirable for a scaffold tree and is
# equivalent to the behavior of SNG (I believe...)
logger.debug(e)
return set()
frags = {Scaffold(f, hash_func) for f in GetMolFrags(frag, True, False)}
return frags
def get_scaffold_frags(frag):
"""Get fragments from a disconnected structure.
Used by fragmentation methods."""
try:
# frag.ClearComputedProps()
# frag.UpdatePropertyCache()
# Chem.GetSymmSSSR(frag)
partial_sanitization(frag)
except ValueError as e:
# This error is caught as dissecting an aromatic ring system,
# may lead to an undefined state where the resultant system
# is no longer aromatic. We make no attempt to prevent this
# but log it for reference.
# This behaviour may be desirable for a scaffold tree and is
# equivalent to the behavior of SNG (I believe...)
logger.debug(e)
return set()
frags = {Scaffold(f) for f in GetMolFrags(frag, True, False)}
return frags
def get_all_murcko_fragments(mol, break_fused_rings=True):
"""Get all possible murcko fragments from a molecule through
recursive removal of peripheral rings
Parameters
----------
mol: rdkit molecule to be processed
break_fused_rings (bool): If True dissect fused rings (default: True)
Returns
-------
A list of rdkit Mols representing all possible murcko fragments
"""
rdlogger.setLevel(4)
if break_fused_rings:
fragmenter = MurckoRingFragmenter()
else:
fragmenter = MurckoRingSystemFragmenter()
mol = get_murcko_scaffold(mol)
rdmolops.RemoveStereochemistry(mol)
scaffold = Scaffold(mol)
parents = {scaffold}
def recursive_generation(child):
for parent in fragmenter.fragment(child):
if parent in parents:
continue
parents.add(parent)
recursive_generation(parent)
recursive_generation(scaffold)
rdlogger.setLevel(3)
return [f.mol for f in parents]
def _contract_rings(mol):
"""Private: Return a molecule with rings contracted.
Create a new molecule, replacing each ring with one
atom. Atoms are connected if the rings they represent
are connected by a bond not in any ring. If rings
share a common bond, the bond added is double. If rings
share an common atom (i.e. spiro rings) the rings are
connected with a single bond.
Used for generating ring connectivity scaffolds.
Parameters
----------
mol : rdkit.Chem.rdchem.Mol
Scaffold template for ring contraction.
Returns
-------
rcs : rdkit.Chem.rdchem.Mol
New molecule with contracted rings.
Notes
-----
Dummy atoms are used instead of carbon atoms, so that
valence constraints are not violated. In ring connectivity
scaffolds the valence of a vertex is occaisionaly > 4.
See Also
--------
get_ring_connectivity_scaffold
"""
# Use a Scaffold object for ring + ring system information.
scf, rcs = Scaffold(mol), RWMol()
# Add a dummy atom for each ring to the empty RWMol.
for _ in range(scf.rings.count):
rcs.AddAtom(Atom(0))
# Create bonds between atoms (ring vertices).
weak_connections, visited = set(), set()
for system in scf.ring_systems:
bc = set(system.get_bond_connectivity())
ac = set(system.get_atom_connectivity())
visited.update(system.aix)
# link strongly connected rings (DOUBLE)
for connection in ac.intersection(bc):
rcs.AddBond(*connection, BondType.DOUBLE)
# link weakly connected rings (SINGLE)
for connection in ac.difference(bc):
rcs.AddBond(*connection, BondType.SINGLE)
# link rings connected by a linker (SINGLE)
for rix, ring in zip(system.rix, system):
to_visit = ring.get_attachment_points()
while to_visit:
aix = to_visit.pop()
for nbr in scf.atoms[aix].GetNeighbors():
idx = nbr.GetIdx()
if idx in visited:
continue
elif nbr.IsInRing():
visited.add(idx)
for nix in _get_rings_with_atom(scf, idx):
c = tuple(sorted((nix, rix)))
weak_connections.add(c)
continue
else:
to_visit.add(idx)
visited.add(idx)
# Add remaining weak ring connections (SINGLE)
for connection in weak_connections:
rcs.AddBond(*connection, BondType.SINGLE)
return rcs.GetMol()