Giters

openforcefield / openff-recharge

An automated framework for generating optimized partial charges for molecules

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

is it possible to allow different tightness parameters?

yanze039 opened this issue · comments

commented

Hi,

I am not sure if I'm cool to post question here. But I notice that the OpenFF RESP documentation website (https://docs.openforcefield.org/projects/recharge/en/stable/users/resp.html) illurstrates that the two-stage resp charge fitting procedure has b=0.1 for both steps,

Any suggestions on how to tune this parameter?
(editted because of misleading...)

commented

I'm completely unqualified to give scientific guidance on this question, but with regards to whether or not you can fiddle with the b parameter: yes! Here's where that parameter is set, and you can use a development build to do these sorts of experiments: https://github.com/openforcefield/openff-recharge/blob/0.5.1/openff/recharge/charges/resp/_resp.py#L411

Something like this, with your conda environment already activated, is how to get a development build rolling:

$ git clone git@github.com:openforcefield/openff-recharge.git
$ cd openff-recharge
$ git checkout origin/0.5.1  # or whatever version you were using
$ vi openff/recharge/charges/resp/_resp.py  # or your preferred text editor/IDE
$ python -m pip install -e .

Allow me to also route the question to @lilyminium who is likely to know more about the wizardry of RESP (but might not be back in the office until next week)

commented

Thanks for the information!

And sorry for my misunderstanding of the materials, the hyperbolic constraints has the form of
$ \chi = a \sum{ ( q^2 + b^2 )^(1/2) -b} $
which really means restraining charges towards 0, no matter what value of b is ...