Re: Force field design

From: Mayne, Christopher G (
Date: Tue Jul 22 2014 - 09:01:58 CDT


I'm not really sure how to elaborate much more than that what I stated in the VMD-L post you linked to. Gaussian is a commercial quantum mechanics software. It is typically used via command line in which you provide it an input file that contains settings for the QM calculation. During the calculation it writes data to a LOG file. As I mentioned in the VMD-L post, ffTK will generate the input file and parse the resulting LOG file. The user is only expected to have access to a Gaussian install and to be able to run the jobs themselves.

Christopher Mayne

On Jul 22, 2014, at 5:40 AM, Abhishek TYAGI wrote:

Dear Mayne,

Thankyou for your detail reply for my query, I had done screencast tutorials, but I faced a problem not getting the output in Step-III for gaussian output log files. If possible could you please tell me how to get these files I read your comment ( I don't know how to use Gaussian, it is new to me, but I will learn it If I knwo how to start.


From: Mayne, Christopher G <<>>
Sent: Sunday, July 20, 2014 8:10 AM
To: Abhishek TYAGI
Subject: Re: Force field design

see my comments below.

On Jul 19, 2014, at 8:31 AM, Abhishek TYAGI wrote:


your fftk tool is good I heard about it, I am new to this thing and trying from last 1 month and I have nothing in my hand, I have following problems:
1. I tried my best to understand it but I find it difficult to do so, I donít understand that I do not have any parameter file still it asked for parameter files.

The very first step of ffTK asks for "Associated Parameter Files" which is used to determine which parameters are required for your molecule that are NOT present in the associated parameter set. This is particularly useful if your molecule includes common substructures that may already be described by a particular force field (e.g., CGenFF). If your molecule is COMPLETELY novel (or you've typed it in such a manner) then you don't have to provide any "Associated Parameter File" and ffTK will generate an initial, zero-ed out parameter file for you to develop as you proceed through the parameterization workflow.

2. I have only my molecule file a.mol2 and a.str (CGenFF) file.

ffTK minimally requires that you have a properly typed PSF and PDB file. There are many ways to generate this from a mol2 and the data contained within the stream file. One such way is using the molefacture plugin in VMD, but there are others.

3. Than I had tried to play with the tool and I get a output parameter file have all the things zero (0), I was correct to this step.... BUT.. next what happen.

yep, see my comment above

4. "Assign Missing VDW/LJ Parameters by Analogy" You mentioned other stuffs also but when I came across the next point it asked for parameter and topology file, If I had them then why I try FFTK.

ffTK doesn't support the parameterization of NONBONDED terms (i.e., vdW). Accordingly we provide a tool that allows you to search through existing parameter sets (parameter + topology files) to assign these terms by analogy.

5. I am really frustrated to know how to get the force field of my a.mol2 file and a.str file.. :(. for your information I am attaching them, might be they are also wrong.

Technically the stream file IS the force field file; you simply need to generate the paired PSF/PDB files to use it. That's a matter of file manipulation and/or using PSFGen. If you want to refine the parameters provided in the stream file, then that is a parameterization question that ffTK is designed to address.

6. One more thing that the pdb and psf have different information and a.str have different, If I try to seperate them than still I cant find any similarity.

PSF, PDB, and STR files contain different information (there is some overlap, actually). You should familiarize yourself with what information you should expect each of those files to have.

thanks in advance

Parameterization is an extremely complex activity. It requires that you understand the basics of the underlying force field and what calculations are requires to derive the parameters. The ffTK documentation provides links to several articles that describe this. I have also posted a set of screencast tutorials that details the steps required, and should be completely sufficient for getting users started. It seems that you also don't have a very good feel for what data is contained in PSF, PDB, STREAM, parameter, and topology files, or how to combine them to construct, setup, and run a simulation. I encourage you to take a look at the VMD and NAMD tutorials, as well as, VMD, NAMD, and PSFGen documentation, to try and resolve these issues prior to attempting parameterization.



From:<> <<>> on behalf of Mayne, Christopher G <<>>
Sent: Wednesday, July 16, 2014 11:46 PM
To: NAMD list
Subject: namd-l: Re: Force field design

As others have pointed out, there are now several resources scattered across the web for helping you obtain force field parameters for a molecule. Some of them work by analogy to molecules for which parameters are already available, others by fitting to first principles calculations, and some provide little more than Frankenstein's monster-type amalgams of parameters from various force fields and/or use derivation methods that are incongruent with the published standards for your target force field. If you are interested in obtaining high quality parameters, you should take care to understand the method the tool uses and whether it is appropriate for the force field employed in your simulations. The proper use of a parameterization tool and massaging the I/O into whatever form you need to make your MD run is really a secondary issue and is more appropriately directed to the authors and/or support systems for that particular program.

I am the author of the Force Field Toolkit (ffTK) distributed within VMD. To obtain parameters compatible with the CHARMM force field, I suggest obtaining initial parameters using the CGenFF Program (formerly Paramchem). This output will include penalty scores that will inform you which terms require additional testing and/or refinement, which can be done using ffTK.

Christopher Mayne

If you are familiar with the CHARMM program, a good resource is:
But for that, of course, you would need very good knowledge of the CHARMM
program (you might even have to even obtain a copy of it which is not
On Tue, 15 Jul 2014, Abhishek TYAGI wrote:
> I want to run MD, but unfortunately I do not have the reasonable force field for this molecule.
> Now what I tried following approaches to get force field:
> 1. I tried CGenFF and got a *.str file but the data in the file is different from my input file i.e. mol2 file, the pdb format is different from the *.str file.
> 2. I tried to get force field from swissparam website and received the different file as *.par, *.top file, I tried MD on this field it runs successful, but the question is that is this correct approach?
> 3. I tried parameter tool available in VMD too I got some output from it too, but I am not able to connect all the information.
> Any guess what to do to get accurate force field parameter for the molecule

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:01 CST