Seminario method or similair

[Erikna00]

Hi Ragnar and thanks for the great work!

Is there any chance that ASH could get a QM based force field parametrization module?
It seems quite ideal to put it in a interface project like this where one could íntegrate every step from optimizing the geometry(ies if several conformers), calculate hessian(s) and fitting a force field for openmm. It seems useful also that in ASH one could reallistically also use the QM optimization to refine the force field in addition to the hessian.

Best, Erikna

[RagnarB83]

Hi Erikna

Thanks for the interest in the ASH project.
Indeed it would be great to to have some kind QM-derived forcefield parameterization option within ASH. I completely agree that it would fit well within the scope of ASH.

Right now, nothing is planned but I will put it on the TODO/EXPLORE list.
The easiest way forward would probably be to write an interface to something that is already established. Maybe that would be QUBEkit or something?

[Erikna00]

Id propose the method by Per-Ola Norrby availible at https://github.com/mmfarrugia/merge-Q2MM

The program is laser focused on gaussian +schrodinger but with some tweaking you could probably branch of a qm and ff agnostic code with little issue.

I propose this mostlly since any time anyones benchmarking the Norrby method always wins and additionally you get a transition state force field parametrization solution "for free".

Nice to hear that the idea was appreciated
Erikna

[RagnarB83]

Great idea, thanks!
I remembers Per-Ola Norrby talking about Q2MM many years ago at a conference. Interesting to hear that it still is the state of the art.

[Erikna00]

I just found that ForceBalance from openMM uses the Norrby method and is currentlly maintained with native support for openMM as well as a fancier optimizer than Q2MM. For spicy molecules (organometalls) a seminario initial guess would need to be implemented and it would be nice if one could write something to let it handle TS force fields too.

So it is a viable alternative to stealing Q2MM with some pros and cons.

[RagnarB83]

Ah yes, I've heard about about Lee-Ping Wang's forcebalance.
I'll take a look at writing an interface to this when I can, probably reasonably straightforward.