Re: [Rdkit-discuss] Exhaustive Library Enumeration

[Christos K. <chr...@gm...>]

Hi Andy,

The reason that your code breaks is that the second product of the third
iteration ( 'NCCCCN(Cc1ccccc1)(Cc1ccccc1)Cc1ccccc1') is not a valid
molecule.
And when calling Chem.MolFromSmiles( 'NCCCCN(Cc1ccccc1)(Cc1ccccc1)Cc1ccccc1')
it creates a None object.
So you have to filter out the molecules that are not valid.

See this Jupyter Notebook
<https://gist.github.com/CKannas/11bb9bcaa9435dd18a0bb969501219b2> at cell
5 the 1st line in the while loop.

Best,

Christos

Christos Kannas

Chem[o]informatics Researcher & Software Developer

[image: View Christos Kannas's profile on LinkedIn]
<http://cy.linkedin.com/in/christoskannas>

On 17 January 2018 at 18:16, Andy Jennings <and...@gm...>
wrote:

> Hi RDKitters,
>
> I have a question and an observation on the topic of library enumeration.
>
> First, the question: is there a call within RDKit to trigger the
> exhaustive reaction of reagents? For example, if I have two reagents - a
> primary amine and an akyl chloride - can I tell RDKit to enumerate the
> reaction as though there were an excess of each reagent? In my case here
> the reaction would continue until the alkylation can no longer occur
> because there are no more valences available on the amine and I would
> either be tri-alkylated for a neutral product or quat-alkylated for a
> positively charged product
> e.g. CCN + RCl -> CCN(R)(R)R or CC[N+](R)(R)(R)R
>
> This brings me to my observation. When I try to attempt exactly this by
> repeatedly exposing the product to the reagent again I am able to drive it
> to exhaustion *in some cases*.
>
> For example, in the example above where RCl is benzyl chloride and my
> smirks is:
> [#7:1].[#6:2][Cl:3]>>[#6:2][#7:1].[Cl:3]'
> I do drive the final product to be exclusively the tri-akylated amine.
> Success.
>
> However, when I attempt the same thing with an amine with more than one
> reactive nitrogen (e.g. NCCCCN) I don't get a single product with 6
> alkylations, I get two unique product each with three alkylations. One
> product has two alkylations on the first nitrogen and one on the second,
> the other product has three alkylations on the first nitrogen and none on
> the second. Attempting to drive the reaction once again leads to a
> 'reaction called with None reactants' ValueError. My dreadful code is below
> and the output is
> Reaction 1: ['NCCCCNCc1ccccc1']
> Reaction 2: ['NCCCCN(Cc1ccccc1)Cc1ccccc1', 'c1ccc(CNCCCCNCc2ccccc2)cc1']
> Reaction 3: ['c1ccc(CNCCCCN(Cc2ccccc2)Cc2ccccc2)cc1',
> 'NCCCCN(Cc1ccccc1)(Cc1ccccc1)Cc1ccccc1']
> Reaction 4: ValueError
>
> Any pointers would be great, as would any pre-existing library enumeration
> code. The examples I've found shipped with RDKit don't appear to allow me
> to name the products using a combination of the reagent names (useful for
> tracking library content).
>
> Best,
> Andy
>
> #### Code snippet ####
>
> amine = Chem.MolFromSmiles('NCCCCN')
> acyl = Chem.MolFromSmiles('c1ccccc1CCl')
> rxn = AllChem.ReactionFromSmarts('[#7:1].[#6:2][Cl:3]>>[#6:2][#7:
> 1].[Cl:3]')
>
> # First reaction
> reactantListMols = [amine,acyl]
> prods = AllChem.EnumerateLibraryFromReaction(rxn,[reactantListMols,
> reactantListMols])
> prods = list(prods)
> smis = list(set([Chem.MolToSmiles(x[0],isomericSmiles=True) for x in
> prods]))
> print smis
> # ['NCCCCNCc1ccccc1']
>
> # Now repeat until doom
> for i in range(0,10):
>     oldproducts = [Chem.MolFromSmiles(x) for x in smis]
>     reactantListMols = oldproducts + [acyl]
>     prods = AllChem.EnumerateLibraryFromReaction(rxn,[reactantListMols,
> reactantListMols])
>     prods = list(prods)
>     smis = list(set([Chem.MolToSmiles(x[0],isomericSmiles=True) for x in
> prods]))
>     print smis
>
> #### End Code ####
>
>
>
> ------------------------------------------------------------
> ------------------
> 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
>
>