Esp grid points

julieborah66 · 2018-09-27 14:04:23

Sir,
I generated esp (in grids) for my molecule of atomic charges and got it into a text file espfit.txt?
Is there a way to know which atom is assigned which grid points, since the text file contains x,y,z coordinate of the grid points and fourth column has esp but how can I know the atom numbers?

sobereva · 2018-09-28 00:39:30

Dear julieborah66,

There is no any correspondence between ESP fitting points and atoms. In principle, ESP value at all points affect charge of all atoms. If you are confused, please check Section 3.9.10 for introduction of CHELPG type of ESP fitting method.

Best regards,

Tian

julieborah66 · 2018-10-08 08:18:02

Hello Sir,
Thanks for your reply, Actually, I am referring to a paper " Charge Anisotropy: Where Atomic Multipoles Matter Most: Christian Kramer, Alexander Spinn, and Klaus R. Liedl, Journal of Chemical Theory and Computation, 2014, 10, 4488−4496" in the page 4490 of statistical analysis they mention a line " Each grid point is assigned to the nearest atom for the analysis " (grid points at a distance of 1.66−2.2 times the vdW radius)

How can I assign in my case if I want to calculate the RMSE of the ESP between the quantum and the CM5 charges. I want not using any fitting CHELP or RESP. It's just that I want to calculate to error metrics.

Last edited by julieborah66 (2018-10-08 08:18:49)

sobereva · 2018-10-08 08:56:54

Please check end of Section 4.7.8 of the latest version of Multiwfn manual, in which I exemplified how to evaluate reproducibility of ESP (measured by RMSE and RRMSE) of given atomic charges. Note that only currently latest version of Multiwfn has this capacity.

julieborah66 · 2018-10-10 11:43:50

Hello Sir,

Apologies to disturb you once again, is there any way out in Multiwfn to calculate the RMSE error around a specific atom?

julieborah66 · 2018-10-10 12:08:57

Sir,

I uploaded a file containing CM5 charges. The phenol.wfn file was taken from the example section and then the CM5 charges were calculated. I uploaded the .chg file in the CHELP module I got RMSE error 0. I also did it with MK scheme. I got RMSE. Why is that so? Is there something wrong in my procedure. I am attaching a screenshot of the results

Last edited by julieborah66 (2018-10-10 12:09:20)

sobereva · 2018-10-10 13:32:36

julieborah66 wrote:

Sir,
I uploaded a file containing CM5 charges. The phenol.wfn file was taken from the example section and then the CM5 charges were calculated. I uploaded the .chg file in the CHELP module I got RMSE error 0. I also did it with MK scheme. I got RMSE. Why is that so? Is there something wrong in my procedure. I am attaching a screenshot of the results
https://i.postimg.cc/VJ82WYVz/SCREENSHOT.png

You didn't use the latest version, the CHELPG and MK interfaces of the latest version of 3.6(dev) on the Multiwfn website look like this

------------ Calculation of CHELPG charges ------------
-3 Toggle using atomic charges in external file, current: No
-2 Toggle loading additional fitting centers from external file, current: No
-1 Toggle using fitting points in external file, current: No
0 Return
1 Start calculation!
2 Set grid spacing, current: 0.567 Bohr ( 0.300 Angstrom)
3 Set box extension, current: 5.291 Bohr ( 2.800 Angstrom)
5 Choose ESP type, current: Nuclear + Electronic
6 Toggle if exporting fitting points with ESP after the task, current: No

You should use a file containing wavefunction information (e.g. wfn, fch) as input file, and then enter this module. Before calculation via option 1, you should first select option -3 to load atomic charges from specified .chg file.

sobereva · 2018-10-10 13:38:48

julieborah66 wrote:

Hello Sir,
Apologies to disturb you once again, is there any way out in Multiwfn to calculate the RMSE error around a specific atom?

There is no such an option. If you really need it, I will consider add such a function in the MK module

julieborah66 · 2018-10-10 16:16:21

Thank you Sir for being so prompt in replying to queries. I do need an option where I can calculate the RMSE error of a specific atom in a given molecule, if I upload the chg file. It will be so useful if such an option will be available in Multiwfn

sobereva · 2018-10-10 20:39:37

julieborah66 wrote:

Thank you Sir for being so prompt in replying to queries. I do need an option where I can calculate the RMSE error of a specific atom in a given molecule, if I upload the chg file. It will be so useful if such an option will be available in Multiwfn

I will update Multiwfn soon, new version will include this option. Once update is finished, I will inform you here.

sobereva · 2018-10-12 19:13:18

julieborah66 wrote:

Thank you Sir for being so prompt in replying to queries. I do need an option where I can calculate the RMSE error of a specific atom in a given molecule, if I upload the chg file. It will be so useful if such an option will be available in Multiwfn

Today I have updated Multiwfn 3.6(dev) on the official site, the function you need has been supported. Please check Section 4.7.8 of the latest version of manual on how to use this function.

julieborah66 · 2018-10-14 03:53:50

Okay Sir, thankyou Sir for updating

Multiwfn forum

#1 2018-09-27 14:04:23

Esp grid points

#2 2018-09-28 00:39:30

Re: Esp grid points

#3 2018-10-08 08:18:02

Re: Esp grid points

#4 2018-10-08 08:56:54

Re: Esp grid points

#5 2018-10-10 11:43:50

Re: Esp grid points

#6 2018-10-10 12:08:57

Re: Esp grid points

#7 2018-10-10 13:32:36

Re: Esp grid points

#8 2018-10-10 13:38:48

Re: Esp grid points

#9 2018-10-10 16:16:21

Re: Esp grid points

#10 2018-10-10 20:39:37

Re: Esp grid points

#11 2018-10-12 19:13:18

Re: Esp grid points

#12 2018-10-14 03:53:50

Re: Esp grid points

Board footer