Re: [Rdkit-discuss] Nitrogen Valence

[Peter S. S. <sh...@gm...>]

Hi,

If the compound is neutral overall and there is a single H where you drew
it, then a valid RDKit SMILES for the nitrogen-containing terminal group is
C[N+](C)(C)[NH-], which is one of the forms I gave earlier.

It is not a zwitterion. Rather, it represents a dative bond. (I am not sure
that all [X+][Y-] bonds are dative bonds, but my guess is that they are.)

Attached are SMILES for some well known nitrogen compounds with adjacent +
and - charges. including nitromethane (lower left). All have single-bonded
"ion pairs", but none are zwitterions. Sorry the drawing (from Slack) is so
small.

Carbon monoxide, [C-]#[0+]. The version of RDKit now hooked up to Slack
can't draw it, but I believe that's due to a known bug that also keeps it
from drawing ethane, CC.

Best,
-P.


On Thu, May 11, 2017 at 1:45 PM, Yuran Wang <wan...@gm...> wrote:

> Hi Peter,
> Thank you for your reply. I did not quite understand what you mean by 'But
> this makes no sense'.
> Also the SMILES you tested are zwitterionic form. In this link
> http://www.rdkit.org/docs/RDKit_Book.html#molecular-sanitization, the
> zwitterionic form seems suitable for N=O, N#N, not for N=N. But I may just
> have a very limited knowledge of RDkit.
>
> This is how it looks like in ChemDraw:
> [image: Inline image 1]
>
>
> Thanks,
> Yuran
>
> On Thu, May 11, 2017 at 1:33 PM, Peter S. Shenkin <sh...@gm...>
> wrote:
>
>> The problematic part is just the beginning of your would-be SMILES:
>> N=N(C)(C)C. The rest is correctly parsed. But this makes no sense. Perhaps
>> you mean one of the substructures illustrated in the attached (which at
>> least satisfy normal valence rules). If not, perhaps you could attach a
>> structural diagram of what you do mean.
>>
>> -P.
>>
>>
>> On Thu, May 11, 2017 at 11:02 AM, Yuran Wang <wan...@gm...>
>> wrote:
>>
>>> Dear Greg,
>>> Thank you very much for the suggestions. It works for me!
>>> Here is the SMILES of one molecule that I am looking
>>> at: N=N(C)(C)CC(CN1N=CN=C1)(O)C2=C(C=C(C=C2)F)F
>>> Any better alternative will be appreciated.
>>>
>>> Thanks,
>>> Yuran
>>>
>>> On Thu, May 11, 2017 at 10:49 AM, Greg Landrum <gre...@gm...>
>>> wrote:
>>>
>>>>
>>>>
>>>> On Thu, May 11, 2017 at 4:24 PM, Yuran Wang <wan...@gm...>
>>>> wrote:
>>>>
>>>>> I have a question regarding the available valence of Nitrogen. It
>>>>> seems only 3 is available in the default setting (atomic_data.cpp). Why is
>>>>> it kept to only 3, and not extended to include 4 and 5? If I change it
>>>>> locally to include 4 and 5, will it cause any problems?
>>>>>
>>>>
>>>> Aside from generating molecules that don't make any chemical sense?
>>>> Probably not, but the lack of chemical sense may cause some unexpected
>>>> behavior.
>>>>
>>>>
>>>>> I am aware that I could turn off the sanitization to get a mol object,
>>>>> however, it cannot be further processed to get fingerprints, which is what
>>>>> I need.
>>>>>
>>>>
>>>> Well, you could turn off the sanitization on molecule construction and
>>>> then manually sanitize with the valence check turned off. Here's a simple
>>>> example of that:
>>>>
>>>> In [11]: m = Chem.MolFromSmiles('CN(C)(C)(C)C',sanitize=False)
>>>>
>>>> In [12]: m.UpdatePropertyCache(strict=False)
>>>>
>>>> In [13]: Chem.SanitizeMol(m,Chem.SANITIZE_SYMMRINGS|Chem.SANITIZE_SET
>>>> CONJUGATION|Chem.SANITIZE_SETHYBRIDIZATION)
>>>> Out[13]: rdkit.Chem.rdmolops.SanitizeFlags.SANITIZE_NONE
>>>>
>>>> In [14]: rdMolDescriptors.GetMorganFingerprint(m,2)
>>>> Out[14]: <rdkit.DataStructs.cDataStructs.UIntSparseIntVect at
>>>> 0x10b0ab350>
>>>>
>>>>
>>>> But, again, the RDKit's valence rules tend to reflect real chemistry.
>>>> What are you trying to represent that you need 5 coordinate neutral
>>>> nitrogen atoms? There may be a better way.
>>>>
>>>> -greg
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Best,
>>> Yuran Wang
>>>
>>> ------------------------------------------------------------
>>> ------------------
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>> _______________________________________________
>>> Rdkit-discuss mailing list
>>> Rdk...@li...
>>> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>>>
>>>
>>
>
>
> --
> Best,
> Yuran Wang
>