On Jun 14, 2018, at 3:07 PM, Noel O'Boyle
<baoille...@gmail.com<mailto:baoille...@gmail.com>> wrote:
Hi Mohammad,
Just to take a step back, could you describe how users of Open Babel will
benefit from this? That is, what they might use these charges for? I know that
Open Babel already has several charge models, but none are documented
currently, and so I have no clue what to do with them except calculate the
values and write them out.
I am parametrizing the EEM to calculate both polarizable and non-polarizable
partial charges for GAFF atom types based on three charge types, namely point
charge, Gaussian, and Slater charges.
We get much better charges than the original EEM methods, almost as good as
CM5, which is the best QM method, because it yields physical charges (in
constrast to ESP) and reproduces the ab-inition dipole moment.
Implementation of the polarizable version in open babel needs much work to do,
but we have it inside GROMACS.
So, we are thinking to implement the non-polarizable version in open babel for
the first step. And for the second step, if possible, to implement the
polarizable version.
A lot of programs widely used in virtual screening like ZINCPharmer and PyRx
can use the code afterward to generate partial charges for a large number of
molecules.
Cheers,
Mohammad
Regards,
- Noel
On 14 June 2018 at 09:36, Mohammad Mehdi Ghahremanpour
<m.ghahremanp...@hotmail.com<mailto:m.ghahremanp...@hotmail.com>> wrote:
Open Babel developers,
We are planning to implement a library of coulomb integrals for Slater- and
Gaussian-type charges into open babel.
We have implemented the Slater integrals (analytical solutions) in two
different versions. The first version uses an arbitrary high precision (using
the CLN library) that supports 1s to 6s Slater orbitals. The second version
uses double precision that only supports 1s to 3s orbitals. The CLN version is
slower than the second version and also makes Open Babel dependent on the CLN
library, but it has a much higher precision and also takes care of 4s and 5s
orbitals that we need for Br and I. However, we have recently parametrized 3s
orbitals for Br and I and our results show that this is a good approximation.
Geoff and I have been discussing whether it is better to implement both
versions and then decide to compile it with or without the CLN library using a
cmake flag or just implement the
double precision version that only supports 1s->3s Slater orbitals which is
more straightforward.
So, what do you think?
Cheers,
Mohammad
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org<http://slashdot.org>! http://sdm.link/slashdot
_______________________________________________
OpenBabel-Devel mailing list
OpenBabel-Devel@lists.sourceforge.net<mailto:OpenBabel-Devel@lists.sourceforge.net>
https://lists.sourceforge.net/lists/listinfo/openbabel-devel
------------------------------------------------------------------------------
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
