On Sat, Jun 1, 2013 at 6:11 AM, Robert DeLisle <[email protected]> wrote:
> You are absolutely correct on all counts. I've also found problems with a
> variety of special cases in which the non-cyclic atom is required for
> aromaticity. Lots of special cases to consider.
>
Oh yeah, if you want the aromaticity to survive the fragmentation and
sanitization, you're going to need to include the doubly bound exocyclic N
or O as an N or O. Here's an example:
In [2]: Chem.CanonSmiles('*=C1NC=CC=C1')
Out[2]: '[*]=C1C=CC=CN1'
In [3]: Chem.CanonSmiles('O=C1NC=CC=C1')
Out[3]: 'O=c1cccc[nH]1'
not a simple story, unfortunately.
I'll integrate your suggestions and do done error checking.
> On May 31, 2013 9:55 PM, "Greg Landrum" <[email protected]> wrote:
>
>> Hi Kirk,
>>
>> Quick FYI: the attachment on the original message didn't make it through.
>> Still, I think the idea is clear enough.
>>
>> A general question I'd have is whether or not you really want to
>> completely ignore the nature of the bond to the non-ring atom. For example,
>> with this scheme you're going to end up getting the same ring systems for
>> quinone and 1,4 cyclohexadiene derivatives:
>> O=C1C=CC(=O)C=C1 -> C1C=CCC=C1
>> CC1C=CC(C)C=C1 -> C1C=CCC=C1
>>
>> Similarly, sulfones in rings will lead to thioethers:
>> O=S1(=O)CCCC1 -> S1CCCC1
>>
>> Is that desirable? If not, you could try adding dummies connected with an
>> appropriate bond type so that you get:
>> O=C1C=CC(=O)C=C1 -> *=C1C=CC(=*)C=C1
>>
>>
>> On Fri, May 31, 2013 at 11:41 PM, Robert DeLisle <[email protected]>wrote:
>>
>>> Maybe the logic is that sophisticated after all - Code Block 4 below.
>>>
>>>
>> The trick with adding the Hs to Ns is, unfortunately, necessary. You
>> should be adding them to aromatic heteroatoms.
>>
>> Code Block 4:
>>> from rdkit import Chem
>>> from rdkit.Chem import AllChem
>>>
>>> sdin = Chem.SDMolSupplier('test.sdf')
>>> sdout = Chem.SDWriter('rings.sdf')
>>>
>>> for m in sdin:
>>>
>>> em = Chem.EditableMol(Chem.Mol())
>>> indexmap = {}
>>>
>>> for a in m.GetAtoms():
>>>
>>> if ( a.IsInRing() ):
>>> indexmap[a.GetIdx()] = em.AddAtom(Chem.Atom(a.GetAtomicNum()))
>>>
>>> for b in m.GetBonds():
>>>
>>> if ( b.IsInRing() ):
>>> em.AddBond(
>>> indexmap[b.GetBeginAtomIdx()],indexmap[b.GetEndAtomIdx()],b.GetBondType() )
>>>
>>> for a in m.GetAtoms():
>>> if ( a.IsInRing() and a.GetAtomicNum() == 7 ):
>>>
>>
>> I would replace the above line with:
>> if a.GetIsAromatic() and a.GetAtomicNum() != 6:
>>
>>
>>> for b in a.GetBonds():
>>> if ( not b.IsInRing() ):
>>> em.AddBond(em.AddAtom(Chem.Atom(1)), indexmap[a.GetIdx()],
>>> Chem.BondType.SINGLE)
>>>
>>> for nm in Chem.GetMolFrags(em.GetMol(), asMols=True):
>>>
>>
>> You probably should be calling either Chem.SanitizeMol() or
>> Chem.RemoveHs() here. I'd recommend the second variant because I don't
>> think you want the Hs in the output.
>>
>>
>>> AllChem.Compute2DCoords(nm)
>>> sdout.write(nm)
>>>
>>
>> -greg
>>
>
------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite
It's a free troubleshooting tool designed for production
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap2
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
