"Every valid SMILES is also a valid SMARTS": I think this is one of John
May's lines, which I was never keen on as it makes people think that if you
treat a SMILES as a SMARTS that it will match the original SMILES. It
mostly will, but I think you have found the difference between the SMILES
and SMARTS treatment of "[2H]" - one means deuterium, the other means an
isotope of mass 2 with a single implicit hydrogen attached. It doesn't
match because the deuterium doesn't have another hydrogen attached. [I
think??]
Regards,
Noel
On Thu, 27 Jun 2024 at 10:05, Wim Dehaen <wimdeh...@gmail.com> wrote:
> I would expect the problem here is kekulization. The SMARTS is pattern
> matching using the kekule structure (i.e. double and single bonds, non
> aromatic atoms) and is not sanitized whereas the SMILES after parsing and
> sanitization has aromatic bonds and aromatic atoms. Try what happens when
> you do a SMARTS match with the SMILES with aromatic atoms:
> `[2H]c1cc([3H])cc(C2=N[C@](C)([37Cl])CC2)c1`
>
> best wishes
> wim
>
> On Thu, Jun 27, 2024 at 10:56 AM pgchem pgchem <pgc...@tuschehund.de>
> wrote:
>
>> Hello all,
>>
>> if every valid SMILES is also a valid SMARTS, why does:
>>
>> select substruct('[2H]C1=CC([3H])=CC(=C1)C1=N[C@](C)([37Cl])CC1'::mol,
>> '[2H]C1=CC([3H])=CC(=C1)C1=N[C@](C)([37Cl])CC1'::mol)
>>
>> yield "True", but:
>>
>> select substruct('[2H]C1=CC([3H])=CC(=C1)C1=N[C@](C)([37Cl])CC1'::mol,
>> '[2H]C1=CC([3H])=CC(=C1)C1=N[C@](C)([37Cl])CC1'::qmol)
>>
>> is "False"? The same is observed when using the @> operator.
>>
>> RDKit 2024.03.3 built from source + PostgreSQL 16.3.
>>
>> best regards
>>
>> Ernst-Georg
>>
>>
>> _______________________________________________
>> Rdkit-discuss mailing list
>> Rdkit-discuss@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>>
> _______________________________________________
> Rdkit-discuss mailing list
> Rdkit-discuss@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>
_______________________________________________
Rdkit-discuss mailing list
Rdkit-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
