[gmx-developers] Contributing new water models

[Lee-Ping Wang]

Hi Mark,

 

Thank you, I appreciate it very much. :)  

 

I understand it’s important not to give users too many options in pdb2gmx that could lead to incorrect results.  Kyle Beauchamp and I are working on a manuscript that tests a grid of available protein force fields in Gromacs vs. several water models, including my contribution.  So far we have found that changing TIP3P -> TIP3P-FB increases protein stability (i.e. decreases alpha-carbon RMSD to the crystal structure) for 15 different proteins, improves NMR J-coupling prediction in over 400 small tripeptides, and improves chemical shift prediction in lysozyme and ubiquitin.  We ran these tests for ff96, ff99sb, ff99sb-ildn and ff99sb-nmr; also we haven’t tested any properties where TIP3P-FB makes the model worse.  This isn’t published yet.

 

If we publish improved results for most of the protein force fields + the new water models, then I might request that the new models be considered for addition to pdb2gmx. We could evaluate whether that decision is appropriate at that future time. Let me know what you think.

 

-          Lee-Ping

 

From: gromacs.org_gmx-developers-bounces at maillist.sys.kth.se [mailto:gromacs.org_gmx-developers-bounces at maillist.sys.kth.se] On Behalf Of Mark Abraham
Sent: Friday, May 30, 2014 12:35 AM
To: Discussion list for GROMACS development
Subject: Re: [gmx-developers] Contributing new water models

 

Hi Lee-Ping,

That sounds useful. Generally we will only incorporate force field things that are published, so that is good. I'm hesitant to make them available with pdb2gmx -water, lest it be too easy for people to misuse them with an inappropriate force field. Incorporating the atom types into the tested force fields and making the .itp files available with them sounds fine to me. Then a quick sed -i will replace whatever was selected with pdb2gmx.

Mark

On May 28, 2014 7:12 PM, "Lee-Ping Wang" <leeping at stanford.edu> wrote:

Dear developers,

I'd like to contribute the following water models that were published in our recent article, "Building force fields - a systematic, automatic and reproducible approach."  The models, called TIP3P-FB and TIP4P-FB, are reparameterized versions of the TIP3P and TIP4P models; they have identical functional forms and computational cost, just different parameters.  They are highly accurate for reproducing the thermodynamic, kinetic and structural properties of liquid water across a wide temperature and pressure range.  The .itp files are attached in their final form.

I've run several microseconds of protein simulations using TIP3P-FB and TIP4P-FB, and they are highly compatible with the AMBER protein force fields (more specifically, amber99sb, amber99sb-ildn and amber99sb-nmr).  In fact, replacing TIP3P with TIP3P-FB tends to improve the stability of the "fast folding" proteins in D.E. Shaw's paper, and it also improves the prediction of NMR J-couplings and chemical shifts.  This is currently unpublished work but I'm working on a manuscript.

Please let me know if there's interest.  I think it could make sense to bundle this with the main Gromacs distribution, because the water models come with atom types that need to go into "ffnonbonded.itp" for the various implemented force fields.

Thanks,

- Lee-Ping Wang

Link to the paper: https://www.dropbox.com/s/phwlr3jsnhe46bw/Wang_Pande.RigidWater.JPCL.2014.pdf




--
Gromacs Developers mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or send a mail to gmx-developers-request at gromacs.org.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20140530/f567a52d/attachment-0001.html>