Re: [Rdkit-discuss] GetMolFrags asMols with persistent atom indexes

[Joshua M. <jos...@be...>]

Hey Again,

Back to this problem because I didn't isolate the problem well enough in
the first place - my bad!

I would like to get the original atom indices of the r groups created by
the ReplaceCore function.
The issue is that the atom indices are reset by the ReplaceCore() function
(and seemingly any other function that deletes atoms).

Any idea how I would get these?

In [1]: from rdkit import Chem

In [2]: from rdkit.Chem.Scaffolds import MurckoScaffold

In [3]: m = Chem.MolFromSmiles('c1cc(C)c(CC(=O)[13C])cc1')

In [4]: murcko = MurckoScaffold.GetScaffoldForMol(m)

In [5]: rGroups = Chem.ReplaceCore(m, murcko, useChirality=True,
labelByIndex=True)

In [6]: rGroup_idxs = []
   ...: rGroups = Chem.GetMolFrags(rGroups, asMols=True,
fragsMolAtomMapping=rGroup_idxs)
   ...: rGroup_idxs
   ...:
Out[6]: [(0, 5), (1, 2, 3, 4, 6)]

I know that the two sets of indices I want from the original molecule m are
((3,2),(5,6,7,8))

(found by depicting the indices - see attached)




Any idea how I can achieve this?

Cheers in advance,
Josh


On 26 June 2018 at 12:25, Paolo Tosco <pao...@gm...> wrote:

> Hi Josh,
>
> if you pass empty lists as the frags and as the fragsMolAtomMapping
> parameters of GetMolFrags() they will be filled with indices referring to
> the indices of the original molecule (see lines highlighted in red below).
>
> Hope that helps, cheers
> p.
> GetMolFrags(...)
>
>
>
> GetMolFrags( (Mol)mol [, (bool)asMols=False [, (bool)sanitizeFrags=True
>     [, (AtomPairsParameters)frags=None [, (AtomPairsParameters)fragsMolAtomMapping=None]]]]) -> tuple :
>     Finds the disconnected fragments from a molecule.
>
>       For example, for the molecule 'CC(=O)[O-].[NH3+]C' GetMolFrags() returns
>       ((0, 1, 2, 3), (4, 5))
>
>       ARGUMENTS:
>
>         - mol: the molecule to use
>         - asMols: (optional) if this is provided and true, the fragments
>           will be returned as molecules instead of atom ids.
>         - sanitizeFrags: (optional) if this is provided and true, the fragments
>           molecules will be sanitized before returning them.        - frags: (optional, defaults to None) if this is provided as an empty list,
>           the result will be mol.GetNumAtoms() long on return and will contain the
>           fragment assignment for each Atom
>         - fragsMolAtomMapping: (optional, defaults to None) if this is provided as
>           an empty list, the result will be a a numFrags long list on return, and
>           each entry will contain the indices of the Atoms in that fragment:
>           [(0, 1, 2, 3), (4, 5)]
>       RETURNS: a tuple of tuples with IDs for the atoms in each fragment
>                or a tuple of molecules.
>
>
> On 26/06/2018 10:29, Joshua Meyers wrote:
>
> Hey All,
>
> I am chopping parts out of molecules and picking up the pieces using the
> Chem.GetMolFrags() function.
> I would like to separate fragments into distinct mol objects but keep the
> original atom indexes however setting the asMols=True flag seems to reset
> the indexes, is this expected behaviour? Is there a way around this?
>
> Here's some code:
>
> In [1]: from rdkit import Chem
>
> In [2]: m = Chem.MolFromSmiles('c1ccccc1.CCO')
>
> In [3]: Chem.GetMolFrags(m)
> Out[3]: ((0, 1, 2, 3, 4, 5), (6, 7, 8)) <----- desired indexes
>
> In [4]: frags = Chem.GetMolFrags(m, asMols=True)
>
> In [5]: for frag in frags:
>    ...:     print([a.GetIdx() for a in frag.GetAtoms()])
>    ...:
> [0, 1, 2, 3, 4, 5]
> [0, 1, 2] < ------ indexes have been reset
>
>
> Many thanks in advance,
> Josh
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot <https://protect-eu.mimecast.com/s/s6W1CX6DYIXlN6U6Hc4j>
>
>
>
> _______________________________________________
> Rdkit-discuss mailing lis...@li...https://lists.sourceforge.net/lists/listinfo/rdkit-discuss <https://protect-eu.mimecast.com/s/tLq6CY6EVILKZ0fGKfEC>
>
>
>