On 29 May 2012 15:25, Geoff Hutchison <ge...@geoffhutchison.net> wrote:
> Some of the generated conformers look really ugly (in terms of close
> atoms/steric clashes).
> I have placed a selected subset at: http://imgur.com/a/LgJqL
>
> Is that to be expected in RMSD scoring -- because I have not used --score
> energy? (I tried the energy scoring, and the molecules look more
> reasonable). Is rmsd scoring a sensible default then? Should these
> molecules be removed by the algorithm in rmsd scoring?
>
>
> Well, there is, by default, a steric filter to eliminate clashes, although
> I see the default cutoff is 1.0 Angstrom.
>
> http://openbabel.org/api/2.3/classOpenBabel_1_1OBStericConformerFilter.shtml
>
> Could you give me some rough ideas on the distances of clashes in your
> images. We can certainly change the default.
>
>
Sure, here they are http://imgur.com/a/DaEKo#1 If you want, I can pass you
the generated file via email.
Perhaps the steric clash filter is only considering heavy atoms, as I can
only find Hs which break the 1A cut-off. I still think that (default)
value is a bit on the low side. Perhaps naively, I would expect the atoms
to be at least their VdV radii far from each other. So the C to C atomic
distance to be at least 1.4A away (perhaps even more). And these
constraints would be atom-type dependant. What do you think?
This value of 1A in the steric filter is not available from the command
line, correct?
> This is a good suggestion. I wonder if there's a good way to do a spelling
> suggestion as well? Let me think about that...
>
>
:-)
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
OpenBabel-discuss mailing list
OpenBabel-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
