Universal Force Field / QEq method

General Discussion Documentation
1

Dear Avodadro team,

Currently I am performing a review about atomic interactions in molecular mechanics and I am looking for the models of any kind of interactions.

I have found the charge equilibration method to obtain atomic charges, developed by Rappé and Goddard (https://pubs.acs.org/doi/abs/10.1021/j100161a070), who published the Universal Force Field, programmed by you.

I would like to know where did you find the set of parameters for that force-field. Also, I want to know if you applied the charge equilibritation method to compute atomic charges. In that case, how did you solve the Coulomb integrals mentioned by Rappé and Goddard in their publication? They do not say how to solve them neither give the atomic parameters for all elements.

Best regards!!

2

For the UFF force field, I re-used the parameters from RDKit, although the parameter files are available from multiple sources:

The UFF method does not use the QEq charges. There are a number of sources out there for this fact, but the best discussion comes from MCCS Towhee

The UFF parameters were derived without the use of point charges on the atoms and I believe the consensus of the original authors is to use this force field without any additional charges. One notable proponent of not using partial charges for UFF is A.K. Rappe who states this quite strongly in his UFF FAQ (no longer available).

Your question is about the Coulomb integrals in QEq. These are Coulomb integrals using Slater-type orbitals and there are plenty of references about this. An open source implementation of QEq is available in Open Babel:

That implementation re-casts the Slater-type integrals into approximate Gaussian integrals with fairly reasonable accuracy.

Hope that helps!

3

Thank you for your reply!!

It is very useful!!

4

Dear Prof. Hutchison,

I have read the documentation in OpenBabel. I have understood that your code uses Gaussian-type orbitals…

Now, I would like to know where did you find these parameters:

https://github.com/openbabel/openbabel/blob/master/data/qeq.txt ??

I would like to describe this in my review.

5

You’re correct - the Open Babel implementation of QEq was contributed by Jiahao Chen who wrote the QTPIE model.
https://doi.org/10.1016/j.cplett.2007.02.065

As you can see from git blame he contributed the qeq.txt parameters:

6

Thank you!!

Would you share an e-mail direction to contact him? Please.

I tried to write him at jiahao@mit.edu but I get this message:

Final-Recipient: rfc822; jiahao@mit.edu
Status: 5.1.1
Action: failed
Last-Attempt-Date: Wed, 09 May 2018 01:32:58 -0400
Diagnostic-Code: smtp; 550 5.1.1 <jiahao@mit.edu>… User unknown

7

I replied by e-mail but for the sake of completeness I’ll reply in case someone finds this here.

I don’t have a newer e-mail address for Jiahao, but he’s on Twitter:
@acidflask

8

Hello mdiaz,

please, do you know the source of parameters in open-babel QEq? Electronegativity and hardness are from the original publication, but I cannot find a source of screening radius. I contacted Mr Chen on Twitter, but he doesn’t reply to me.

Best regards!

9

Dear ondrej.schindler. I tried to contact him two years ago, but I never received answer. I don’t know what the parameters source is. If you achieve to contact him or get the parameters source it wuld be a good idea to share it and attach the info in the qeq.txt file of OpenBabel poroject.

10

My guess - having talked with him as he contributed the code to Open Babel - is that these were fitted to reproduce the Slater-type parameters in UFF.

Some of the UFF / QEq parameters, sadly, were never published - only incorporated into code in the Goddard group.

11

Thank you for your responses! Unfortunately, I didn’t receive a response too. Probably, as ghutchis wrote, these parameters were not published.