I'm going to edit DeleteAtom so that any Stereo object containing the
deleted atom is invalidated but all other stereos remain unchanged.
For the record, deleting all of the stereo objects is the wrong
solution as the stereo cannot be lazily reperceived (in the general
case). It was known at the point of reading the molecule in and forms
a part of the 3D structure; if deleted, there's no way to work out
after the fact whether a particular chiral centre was specified or
unknown, clockwise or anticlockwise (with the exception of this latter
for 3D structures).
- Noel
On 1 May 2013 22:10, Noel O'Boyle <baoille...@gmail.com> wrote:
> Using Chris's op, and a shorter testcase ("C=CC1=CN=CC1"), the problem
> is that when you delete an Atom, the perceived stereo is invalid
> (well, may be invalid). To work around this, you should remove the
> perceived stereo. We should probably sort it out at our end.
>
> In C++, I have:
> while (pmol->DeleteData("StereoData"));
>
> The equivalent Python might be:
> success = pmol.DeleteData("StereoData")
> while success:
> success = pmol.DeleteData("StereoData")
>
> You shouldn't have to do this immediately after deleting each single
> atom, but only once you are done modifying the structure.
>
> - Noel
>
> On 25 April 2013 16:33, Chris Morley <c.mor...@gaseq.co.uk> wrote:
>> On 24/04/2013 16:01, Pascal Muller wrote:
>>>
>>> Hi,
>>>
>>> I have a little pybel script for generating Murcko scaffolds, which
>>> works well with 2.3.1, but fails on a few molecules ("Segmentation
>>> fault") with 2.3.2.
>>>
>>> Depending on code modifications, the fault raises at different points,
>>> but I have no clue on why the bug is happening.
>>>
>>> Below is my code shortened as much as I could.
>>>
>>> Regards,
>>> Pascal
>>>
>>>
>>>
>>> #!/usr/bin/env python
>>> # encoding: utf-8
>>> import sys
>>> import os, pybel
>>> import openbabel as ob
>>>
>>> def GetMurckoFramework(mol):
>>> haveAtomWithOneNeighbor,atomsWithOneNeighborList =
>>> checkIfIsOnlyOneNeighbor(mol)
>>> while (haveAtomWithOneNeighbor):
>>> for atomToDel in atomsWithOneNeighborList:
>>> mol.OBMol.DeleteAtom(atomToDel) # delete atom which are
>>> not part of a cycle or a linker (terminal atoms)
>>> haveAtomWithOneNeighbor,atomsWithOneNeighborList =
>>> checkIfIsOnlyOneNeighbor(mol)
>>> return mol
>>>
>>> def checkIfIsOnlyOneNeighbor(mol):
>>> isAnAtomWithOneNeighbor = 0
>>> atomWithOneNeighborList = []
>>> for atom in ob.OBMolAtomIter(mol.OBMol):
>>> if not atom.IsInRing():
>>> neighborsCount = 0
>>> for neighbor in ob.OBAtomAtomIter(atom):
>>> neighborsCount = neighborsCount + 1
>>> if (neighborsCount <= 1): # It is a terminal atom (only
>>> one neighbor) double bond
>>> atomWithOneNeighborList.append(atom) # will be deleted
>>> isAnAtomWithOneNeighbor = 1
>>> return isAnAtomWithOneNeighbor, atomWithOneNeighborList
>>>
>>>
>>> def main():
>>> if len(sys.argv) < 2:
>>> print "No input file provided: Murcko.py filetosprocess.smi"
>>> sys.exit(1)
>>>
>>> mol1 = pybel.readstring("smi", "C1(=Cc2cn(nc2)CC(=O)N)C(=O)NC(=O)S1")
>>> mol2 = pybel.readstring("smi",
>>> "C1(=CC2=C(N=C(C2)C(=O)N)C)C(=O)NC(=O)S1")
>>> li = [mol1, mol2]
>>>
>>> for mol in li:
>>> mol = GetMurckoFramework(mol)
>>> print mol
>>>
>>> if __name__ == '__main__':
>>> main()
>>
>>
>> The problem seems to be in the stereo perception code called during smiles
>> output. It possibly arises because the atom deletion somehow messes up the
>> molecule and the stereo perception routines don't cope. I've tried deleting
>> the stereochemical data in the scaffold to little effect, but most effective
>> was to suppress stereo perception in smilesformat output. The output option
>> -xi (C++ pConv->AddOption("i");) nearly does this but needs line 3608 in
>> smilesformat.cpp to be
>> bool is_chiral = isomeric && AtomIsChiral(atom);
>>
>> But this is only a partial solution and the stereo perception needs to be
>> more robust. The bigger problem is how to handle OBMol objects sometimes as
>> patterns rather than as molecules.
>>
>> Remember also that atom indices can be invalidated when another atom is
>> deleted, although that doesn't seem to be an issue in your code.
>>
>> Attached is an op based on your method used from obabel (or the gui) for
>> debugging.
>>
>> Chris
>>
>>
>> ------------------------------------------------------------------------------
>> Try New Relic Now & We'll Send You this Cool Shirt
>> New Relic is the only SaaS-based application performance monitoring service
>> that delivers powerful full stack analytics. Optimize and monitor your
>> browser, app, & servers with just a few lines of code. Try New Relic
>> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
>> _______________________________________________
>> OpenBabel-discuss mailing list
>> OpenBabel-discuss@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
>>
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and
their applications. This 200-page book is written by three acclaimed
leaders in the field. The early access version is available now.
Download your free book today! http://p.sf.net/sfu/neotech_d2d_may
_______________________________________________
OpenBabel-discuss mailing list
OpenBabel-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
