Paul already mentioned an approach using the molecule's RingInfo structure,
which you could certainly use to see if the RDKit thinks the two molecules are
part of the same smallest ring.I suspect that you might want something a bit
more general (apologies that I don't have a copy of that paper handy, so I
can't read what the Roche folks used).
Fortunately, the more general case actually turns out to be easier.I believe
that what you want is to just determine whether or not the path between the two
atoms is composed solely of ring bonds.You could express this in SMARTS
as:{DONOR}@*@*@{ACCEPTOR}That's the three-bond separation case.Where {DONOR}
and {ACCEPTOR} are the SMARTS definitions you would like to use for those
atoms. The "@" in SMARTS means "Ring bond", so this pattern corresponds to a
donor separated by three ring bonds from an acceptor.
This will, unfortunately, also match the case where the donor and acceptor are
adjacent to each other in a 5 membered ring. If it's important than you not
match that, you'll have to do a bit more work.
Best,-greg
_____________________________
From: Marco Stenta <[email protected]>
Sent: Thursday, September 29, 2016 4:13 PM
Subject: [Rdkit-discuss] Ring membership
To: <[email protected]>
Dear Colleagues,I am working on intramolecular hydrogen bonds, formed between
donor-acceptors separated by 2,3,4 atoms, as described in the Roche paper:Kuhn,
B.; Mohr, P.; Stahl, M. J. Med. Chem. 2010, 53, 2601.
I will use SMARTS and RDKIT to identify and classify IMHB.
I see from documentation how to identify atoms that are members of a ring, how
can I identify if two atoms are member of the same ring?
Example:
Donor------X-----Y-------Acceptor
NCC=O
This will match C1CC(NC1)C=O
but I would lie to flag if two or more atoms belong to the same ring, as this
will impact the capability of IMHB to opne and close.
Do you have any suggestion?
cheers,Marco
------------------------------------------------------------------------------
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss