VMD-L Mailing List

From: Marinus Veldhuizen (Mac.Veldhuizen_at_uantwerpen.be)
Date: Mon Oct 07 2024 - 09:30:35 CDT

Dear JC,

Thank you very much for your input and suggestion. Indeed, the paper you're referring to is nice, and I will take the distribution of the cis-/trans conformation to calculate the final charge.
Your input definitely helped to point me into the right direction, the d_QMs - d_MMs were odd indeed. And I found that the Lennard-Jones (LJ) parameters I initially assigned were incorrect. I assigned them by analogy using the CGenFF tool, but the unconventional nature of the molecule needed a more manual approach.

To anyone encountering the same issue. I recommend printing the Lennard-Jones (LJ) energy values and partial charges separately at each optimization step.
Conveniently, there is a line in the fftk_ChargeOpt.tcl file which will do this for you. You only have to uncomment (remove the #) the line;
 puts "($currTraj,$currFrame) $totalEele \t $totalElj \t $totalE"
Which is in line 1084 in VMD1.9.4a57, and it will show you the interaction energies during the optimization.
In my case, it showed a high final LJ potential energy for an atom, indicating a poor reflection of reality.

Thanks a lot for your input,
Mac

________________________________
From: Gumbart, JC <gumbart_at_physics.gatech.edu>
Sent: Sunday, October 6, 2024 2:26 PM
To: Marinus Veldhuizen <Mac.Veldhuizen_at_uantwerpen.be>
Cc: vmd-l_at_ks.uiuc.edu <vmd-l_at_ks.uiuc.edu>
Subject: Re: vmd-l: Re: FFTK partial charge assignment weak acid ONOH (HNO2) - Water Shift settings

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi Mac,

You might take a look at this paper, which parameterized HNO2 for the GROMOS force field: https://urldefense.com/v3/__https://pubs.acs.org/doi/10.1021/acs.jpcb.9b08172__;!!DZ3fjg!5PcijDbdXNJJ9bBaQyig6CA56B7ZGy5kqNTcdSC-YiMzM0-GfTqZwa7XCU1no_eVZD-RHLNWOVkyA602qL5LUKCrZ83rwaB9$
While not exactly what you are doing, it should be informative. Indeed, your more realistic charges look to be similar.

So why does the water shift end have to be increased so much? I don’t know. Is the d_QM - d_MM ultimately reasonable? Or is it also pretty big? You can also use the Charge Optimization Log Plotter to dissect the interactions and see which are more problematic.

Best,
JC

On Oct 3, 2024, at 8:44 AM, Marinus Veldhuizen <Mac.Veldhuizen_at_uantwerpen.be> wrote:

You don't often get email from mac.veldhuizen_at_uantwerpen.be<mailto:mac.veldhuizen_at_uantwerpen.be>. Learn why this is important<https://urldefense.com/v3/__https://aka.ms/LearnAboutSenderIdentification__;!!DZ3fjg!5PcijDbdXNJJ9bBaQyig6CA56B7ZGy5kqNTcdSC-YiMzM0-GfTqZwa7XCU1no_eVZD-RHLNWOVkyA602qL5LUKCrZ9BqmGIl$ >
Dear all,

I dug more in the forum and found two topics that answered some of the questions. Named "FFTK charge optimisation does not converge" & "fftk - MM/QM distances not converging".
They state that the mismatch between QM/MM indicates an issue with your system. And, the objective value during the charge optimization is incredibly high for my molecule, caused by the interaction energy. Hence, there is possibly something wrong with my system.

I am sure the QM geometric optimization is copied correctly from the Gaussian input. I also checked that all calculations in Gaussian were completed successfully. Additionally, I tried a higher accuracy QM method (even though that is not in line with the CHARMM philosophy). Unfortunately, that did not help.

Therefore, I was wondering if anyone has any suggestions on what could be wrong? As I am out of ideas.

Thanks a mil,
Mac
________________________________
From: owner-vmd-l_at_ks.uiuc.edu<mailto:owner-vmd-l_at_ks.uiuc.edu> <owner-vmd-l_at_ks.uiuc.edu<mailto:owner-vmd-l_at_ks.uiuc.edu>> on behalf of Marinus Veldhuizen <Mac.Veldhuizen_at_uantwerpen.be<mailto:Mac.Veldhuizen_at_uantwerpen.be>>
Sent: Wednesday, October 2, 2024 10:15 AM
To: vmd-l_at_ks.uiuc.edu<mailto:vmd-l_at_ks.uiuc.edu> <vmd-l_at_ks.uiuc.edu<mailto:vmd-l_at_ks.uiuc.edu>>
Subject: vmd-l: FFTK partial charge assignment weak acid ONOH (HNO2) - Water Shift settings

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Dear VMD community member,

Thank you for making time to read this email. I am currently working on calculating the partial atomic charges for the ONOH molecule using the Force Field Toolkit (FFTK) in VMD and have encountered some challenges. The initial partial charges obtained with the standard settings appear unrealistic for both cis and trans conformation. However, after adjusting the "Water Shift - End" setting under the "Advanced" tab, and increasing it to 3.0 (from the default value of 0.4), I observed that the partial charges are going towards a more sensible value.

Although the partial charges look more sensible now, given that the final partial charge should be somewhere between the charges of the two conformations, I am a bit wary when doing parameterization. Therefore, I was wondering if changing the distance of the water this much is sensible. Hence, I have the following questions you can hopefully shed some light on.

  *
Is the initial charge (with the standard settings) unrealistic because there is a mismatch between the QM data and the MM data? Because by changing the distance a better solution can be found for ΔU MM-QM to Δd MM-QM [1]? And therefore the solution is no longer found at ΔU = Δd = 0?
([1] Introduction to Force Field Toolkit - Christopher Mayne / youtube around 14:30 )
  * Would increasing the water shift distance significantly cause the QM-calculated input data for the charge calculated to be not applicable anymore? (Because the QM data is calculated at a much shorter distance)
  *
Could it be that the issue of the partial charges arises because it is an acid (pKa = 3,398)?
     *
Because of the influence of water on the acid, FFTK has difficulty to assessing the polarization (μMM-μQM)?
  *
Lastly, if the water shift is not an option, do you have any advice on how to proceed?

I tried searching in the literature but could not find an explanation. Additionally, I could not find an email which had a similar issue. Therefore, hopefully, you can give me some input. Sorry for the long mail.

Kind regards,
Mac

P.S. Here is some additional information

For the trans conformation, the partial charges were calculated as:
standard settings                     O: -0.996 - N: 0.995, O: -0.413, H: 0.414
increased "Water Shift - End"           O: -0.229 - N: 0.194, O: -0.398, H: 0.433
while for the cis conformation, the values were:
standard settings                     O: -0.924, N: 0.083, O: -0.134, H: 1.000
increased "Water Shift - End"           O: -0.505, N: 0.378, O: -0.689, H: 0.816

This might not be of use but I have tried the following:
Doubled checked the geometry.
Set All z-coordinates of all atoms to 0 (perfectly planar molecule).
Increased initial water positions by 50%/100% (pre-QM calculations).
Compute the esp charges (using Gaussian) to see if there is something wrong. It shows that the influence of water does not cause abnormal esp charges.
Different versions of VMD, 1.9.3 and 1.9.4a57
I use Gaussian 16

I have parameterized other molecules successfully, so I do not think there is something wrong with my VMD. Finally, I tried to add the files of the computation but that is not allowed?