Topic Options
#24441 - 05/27/10 05:50 AM Web server to generate CHARMM topology/parameters
VZoete Offline
Forum Member

Registered: 09/05/05
Posts: 8
Loc: Switzerland, Lausanne
Dear CHARMM users,

The Molecular modeling Group of the Swiss Institute of Bioinformatics is pleased to inform you that the SwissParam web server is now open:

This service provides topologies and parameters for small organic molecules compatible with the CHARMM all atoms force field, for use with CHARMM and GROMACS.

The server is fully automatic, and the user only needs to provide the organic molecule of interest in the mol2 format.
The mol2 file must contain all atoms, including hydrogens, and correct 3D coordinates.
Topology and parameter files are sent to the user by e-mail, generally within minutes.

The parameters are derived from the Merck Molecular ForceField (MMFF). Dihedral angle terms are taken as is, while only the harmonic part of the bond, angle and improper terms are retained. Charges are taken from MMFF. Van der Waals parameters are taken from the closest atom type in CHARMM22.

This service is free for academic use. A CHARMm license is required for users from private companies.

The SwissParam approach to derive rapidly topologies and parameters for small organic molecules is also used by the SwissDock docking web service that we opened recently ( SwissDock uses EADock, our docking software based on the CHARMM force field:
The success rate of the docking approach, using the SwissParam parameters, is close to 75 %, illustrating the relevance SwissParam for such applications.

Please, note that, although SwissParam is working properly, the service is in beta version.
In case of problem, or if you'd like to send us some feedback, don't hesitate to contact us.

Vincent Zoete

#24554 - 06/10/10 05:05 PM Re: Web server to generate CHARMM topology/parameters [Re: VZoete]
rmv Online   content

Forum Member

Registered: 09/17/03
Posts: 8419
Loc: 39 03 48 N, 77 06 54 W
I have my doubts about the notion that MMFF parameters are CHARMM compatible. While the CHARMM program supports the MMFF parameters historically, mixing MMFF and CHARMM parameters has never been a focus, or recommended. Using the CHARMM L-J parameters with the MMFF dihedrals seems like a particularly bad idea. Using the MMFF charges would also seem to be a source of divergence from CHARMM compatibility. Consequently, I have some concerns about the quality of parameters generated by this approach, esp. for more detailed studies beyond simple docking.
Rick Venable
computational chemist

#24575 - 06/14/10 10:18 AM Re: Web server to generate CHARMM topology/paramet [Re: rmv]
VZoete Offline
Forum Member

Registered: 09/05/05
Posts: 8
Loc: Switzerland, Lausanne
Dear Rick,

Thank you for mentioning these concerns.

Although we explicitly list the applications where SwissParam was tested in other posts ( and
the situation might not appear clearly enough in this particular announcement.

This is therefore a nice opportunity to underline that SwissParam should not be taken nor used for what it is not intended for.

SwissParam has been developed in the context of rapid Drug Design calculations, generally based on static structures (i.e. docking, minimization, LIECE, MM-GBSA), for which we obtained satisfactory results.
SwissParam has proven in our hands to be very useful in this particular field for obtaining immediately topologies and parameters for a very large number of whatever organic molecules.
SwissParam gives MMFF-derived parameters mathematically compatible with the CHARMM force field equation. It generally provided a depiction of the equilibrium conformers and of the polar functions that appeared sufficient for such applications, which do not require the finest description of the dynamical behavior of the small molecules.

We would like however to emphasize that SwissParam has not been tested in more demanding applications where a very fine description of the molecule is needed (e.g. free energy perturbationů). SwissParam has indeed not been developed for this purpose. In view of the approximations you mention in your post, we clearly recommend optimizing the parameters and/or using the CGenFF for this. The SwissParam rtf/par files could then be used as a starting guess (among others) for CGenFF.

In summary, SwissParam has been developed for rapid Drug Design calculations; some examples being listed above. Other applications should be made with great caution.

Have a nice day,

#26032 - 12/20/10 10:08 PM Re: Web server to generate CHARMM topology/paramet [Re: VZoete]
Kenno Offline
Forum Member

Registered: 12/19/05
Posts: 1535
Loc: Baltimore, MD
Considering rmv's pertinent reservations and VZoete's helpful and honest clarification, I think it would be safer to use the newly released CGenFF program for more detailed studies beyond simple docking, such as MD simulations. And even when using the CGenFF program, it is strongly advised to look at the penalty scores and re-optimize parts of the molecule with high penalties. And even when the penalties are low, a minimal amount of validation and common sense is required. wink

Edited by Kenno (12/20/10 10:15 PM)
Edit Reason: No HTML preview

#37177 - 10/23/18 02:16 AM Re: Web server to generate CHARMM topology/paramet [Re: VZoete]
soroush Offline
Forum Member

Registered: 10/11/18
Posts: 3
Loc: iran
which kind of format that generated with swissparam is related to topology of structure? .itp or .rtf ??


Moderator:  alex, chmgr, John Legato