Hi Diogo,
The easiest way to do this is to use the substructure matching code with
"uniquify=False" to find all the automorphisms between a molecule and
itself:
In [8]: m1 = Chem.MolFromSmiles('Oc1ccccc1')
In [9]: list(m1.GetSubstructMatches(m1,uniquify=False))
Out[9]: [(0, 1, 2, 3, 4, 5, 6), (0, 1, 6, 5, 4, 3, 2)]
Here's another example:
In [10]: m = Chem.MolFromSmiles('Oc1ccc(c2ccc(Cl)cc2)cc1')
In [11]: list(m.GetSubstructMatches(m,uniquify=False))
Out[11]:
[(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13),
(0, 1, 2, 3, 4, 5, 11, 10, 8, 9, 7, 6, 12, 13),
(0, 1, 13, 12, 4, 5, 6, 7, 8, 9, 10, 11, 3, 2),
(0, 1, 13, 12, 4, 5, 11, 10, 8, 9, 7, 6, 3, 2)]
I hope this helps,
-greg
On Sat, Apr 16, 2022 at 1:46 AM Diogo Martins <diogo.stm...@gmail.com>
wrote:
> Hello,
>
> I'd like to enumerate all possible permutations of symmetric atoms.
> Consider the following code:
>
> phenol = Chem.MolFromSmiles("Oc1ccccc1")
> equivalencies = list(Chem.CanonicalRankAtoms(mol, breakTies=False))
> print(equivalencies)
> [0, 6, 4, 2, 1, 2, 4]
>
> Atoms that have the same value in list "equivalencies" are symmetric. For
> phenol, the equivalent atoms correspond to a 180 degree rotation of the
> aromatic ring over the axis containing the carbon-oxygen bond. The possible
> permutations, expressed as atom indices, are:
> [0, 1, 2, 3, 4, 5, 6]
> [0, 1, 6, 5, 4, 3, 2]
>
> By permutations, I mean that it is possible to replace the coordinates of
> the atoms and produce a realistic molecule.
>
> A brute force approach comes to mind, where one would enumerate all
> possible combinations, and exclude those that change the molecular graph.
> In the example above there are four possible combinations, because there
> are two groups of two symmetric atoms. An example of an "invalid"
> combination is swapping the third and seventh atoms without swapping the
> fourth and sixth atoms:
> [0, 1, 6, 3, 4, 5, 2]
> This would be excluded as it breaks the bond between the third and fourth
> atoms (among other bonds).
>
> Is there a method in the RDKit to enumerate the valid permutations?
>
> Thank you,
> Diogo
>
> _______________________________________________
> 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
