Hi,
I agree with what has been said so far, and the key point for me is to have a
tool that does what's expected.
If the method used to generate the initial conformation creates reasonable bond
angles and distances, then no minimization should be performed.
Although, if that doesn't happen, geometry must be optimized (even if with 10
steps or so) and that should be the default.
In other words, there is no point in generating structures that could be easily
identified as 'wrong' before being written. A way to deal with this might be to
do a single point energy evaluation, to check if the energy is above a given
threshold (which can be slightly positive, for that matter) and then run the
minimization.
I agree that torsions are not the top priority, but the minimization should
also take care of bad geometries where two or more molecule branches overlap
(which is probably more relevant for small organic and biological structures).
S
On 06/14/2018 10:58 PM, David van der Spoel wrote:
Den 2018-06-15 kl. 01:20, skrev Geoffrey Hutchison:
Hi,
One of the GSoC projects this summer is an improved fragment-based coordinate
generation method. While that's going, I've been reading a lot of papers on 3D
coordinate generation.
As near as I can tell, no other program attempts to do a conformer search as
part of the 3D generation. Everyone just generates one 3D geometry and then
lets people do conformer searching on top if they choose.
Some of the papers (e.g. Pierre Baldi) compare to Open Babel and then say that
babel is slow - but that's because a lot more work is performed.
I'd like to change --gen3d to default to creating a geometry. Optionally,
geometry optimization can be performed, but the conformer searching will be
separate:
--gen3d # default: will only create a geometry
--gen3d=opt # will perform MMFF or UFF optimization on the geometry
I think for many uses, this will provide much faster performance.
From your description above it is the conformer searching that makes it slow,
so why not make that optional rather than the optimization?
In addition it would be nice if the 3D generation would not crash every once in
a while (at least it does in Avogadro). I guess due to numerical issues.
Thoughts? Concerns?
-Geoff
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
OpenBabel-Devel mailing list
OpenBabel-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openbabel-devel
--
Stefano Forli, PhD
Assistant Professor
Dept. of Integrative Structural
and Computational Biology, MB-112A
The Scripps Research Institute
10550 North Torrey Pines Road
La Jolla, CA 92037-1000, USA.
tel: +1 (858)784-2055
fax: +1 (858)784-2860
email: fo...@scripps.edu
http://www.scripps.edu/~forli/
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
OpenBabel-Devel mailing list
OpenBabel-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openbabel-devel
