VMD-L Mailing List

From: Gumbart, JC (gumbart_at_physics.gatech.edu)
Date: Thu Feb 06 2025 - 22:12:19 CST

Hi Harini,

I’ve taken a look at the files now and was able to reproduce the crash. The issue is that (at least one) of your waters creates a steric clash - specifically that for atom H6 (a few others like H7 don’t look great either). You can see this in the log file and also by visualizing that interaction in VMD.

H5 QME: -8.388 MME: -7.387 (1.000) QMD: 2.312 MMDistDelta: 0.400
H6 QME: 1026271.434 MME: 140917851.092 (139891579.658) QMD: 2.000 MMDistDelta: -0.400
H7 QME: 12.405 MME: 965.773 (953.368) QMD: 2.164 MMDistDelta: 0.400
H8 QME: -9.104 MME: -4.624 (4.480) QMD: 2.278 MMDistDelta: 0.400

While FFTK automates a lot of steps, it can’t be used blindly. My recommendation would be to force the charge of H6 to match H24 (both being bonded to a nitrogen) and then weight the interaction of H24 with a water 2x. That will probably address this problem sufficiently, but you’ll likely have at least one more troublesome one.

One more thing I’ll say: I probably wouldn’t bother to parameterize most of this molecule from scratch. We typically first run it through the CGenFF server and only worry about parameterizing those charges/parameters with a penalty above 10. For example, in your molecule, you have a lot of non-polar methyl and methylene groups. In the CHARMM force field, these are almost always assigned 0.09 for the hydrogens and -0.18 or -0.27 for the carbon (to make the group neutral). There’s rarely any benefit to deviating from this.

Best,
JC

On Jan 28, 2025, at 8:15 PM, Gumbart, JC <gumbart_at_physics.gatech.edu> wrote:

Hi Harini,

Could you send me the files separately along with exactly what you tried (maybe a screenshot of the settings) so that I can first try to recreate the error?

Best,
JC

On Jan 28, 2025, at 6:19 AM, Harini S H <harinih_at_iisc.ac.in> wrote:

You don't often get email from harinih_at_iisc.ac.in<mailto:harinih_at_iisc.ac.in>. Learn why this is important<https://urldefense.com/v3/__https://aka.ms/LearnAboutSenderIdentification__;!!DZ3fjg!473FhBRwXUb7DSfw9nu6b6jFANf4TbekFWt9y3ohjct0HSpaxGxGVC9uSjr4Snhmff99rBhDfVba2z1sC6ah5xnq9N4$>
Dear all,
          I am using FFTk module in VMD1.9.4a in conjunction with Gaussian 16 RevC to parameterise a positively charged small molecule (40 atoms)
          In the Charge optimisation step, I get the following error:
            "Optimization completed. -- Stationary point found. No Mulliken charges found."
             though the Mulliken charges are listed in the log files of the water-interaction data.
          The only other information I have in the debug file is
            'ffTK has halted on error. The optimizer has returned an empty result, indicating an abnormal termination.'
          From similar queries in the forum, I gathered it might be a version incompatibility issue. Hence, I had tested it with the Gaussian versions available to me (G16 RevC and G09 RevA) along with older VMD versions (1.9.3 and 1.9.2) but to no avail.
          Any help is much appreciated. I can share the log files if required.
Regards,
Harini