Hi Richard,

On Mon, Nov 13, 2017 at 4:01 PM, Richard Marchese Robinson <
r.l.marcheserobin...@leeds.ac.uk> wrote:

>
>
> Firstly, thank you for making such a great tool available to the community
> and for continuing to develop it.
>
>
You're welcome! Thanks for the kind words.


>
>
> I have taken the SMARTS patterns, documented here [
> http://www.rdkit.org/docs/GettingStartedInPython.html#
> feature-definitions-used-in-the-morgan-fingerprints], for defining acidic
> and basic groups. (I appreciate the SMARTS will assign atom types.)
>
>
>
> However, assuming a “basic” group means “protonated at typical pH = 7 –
> 7.4”, I believe there is a mistake in the basic group SMARTS pattern.
>
>
>
> This identifies the nitrogen atoms in the following compounds as basic:
> (1) aniline [c1ccccc1N]; (2) methyl thiazole [C[n+]1cscc1].
>
>
>
> Methyl thiazole does not have an atom which can receive a proton, to the
> best of my knowledge.
>
>
>
> Protonated aniline has a pKa of 4.6 in H2O [http://evans.rc.fas.harvard.
> edu/pdf/evans_pKa_table.pdf].
>
>
>
> Hence, whilst no claim is made to have comprehensively integrated all the
> information from Evan's Table, I suggest adapting the SMARTS as follows:
>
> [$([N;H2&+0][$([C]);!$([C,a](=O))]),$([N;H1&+0]([$([C]);!$([
> C,a](=O))])[$([C]);!$([C,a](=O))]),$([N;H0&+0]([C;!$(C(=O))
> ])([C;!$(C(=O))])[C;!$(C(=O))])]
>
>
>
> Does that sound reasonable?
>

You make valid points. The RDKit is using a simple set of "standard" (well,
at least published) feature definitions. In the end, any pharmacophoric
feature definition is bound to be both partial and, at times, wrong. What's
most important is to be consistent.
Having said that, if you have your own feature definitions that you'd like
to use, the RDKit does make that possible; there's some information on that
here:
http://www.rdkit.org/docs-beta/GettingStartedInPython.html#morgan-fingerprints-circular-fingerprints

Best,
-greg
------------------------------------------------------------------------------
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
Rdkit-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss

Reply via email to