|Posted on Tuesday, February 06, 2007 - 4:59 pm: |
I'd like to plumb WebMO to use GROMACS as one of the computational engines. Is there a HowTo for this?
Post Number: 4
|Posted on Tuesday, February 06, 2007 - 5:15 pm: |
You are certainly welcome to do this; we simply ask that you don't redistribute your modified copy of WebMO. (You can, of course, redistribute a patch file, etc. We also might be interested in incorporating your modifications, with your permission, into the official distribution).
By way of background, let me explain why we have not, as you yet, incorporated support for Gromacs/CHARMM/NAMD or the other popular MD programs. It seems as if the input files for these programs are so complex and specific, that it would be hard to write general 'job templates' for these programs. Also, there is the always tricky issue of picking parameters / assigning atom types to initial geometries. As a result, our thinking was thought it would be difficult to make the input mechanism general enough to be useful. Of course, your thoughts / inputs are welcome with regard to this.
Output visualization is another issue, and I think that in this respect WebMO could be useful. Unfortunately, the quantities of interest MD simulations are generally very specific to the situation at hand, and thus hard to encapsulate in a few simple quantities.
Let me know what you are thinking, and I can advise you on how to best proceed.
|Posted on Monday, May 28, 2007 - 4:16 pm: |
It would be very interesting to see if WebMO went in the direction of supporting MM/MD jobs. I am most familiar with AMBER so I will give my thoughts regarding this program, although I assume most of the ideas would be generally transferable to other MM/MD programs (CHARMM, GROMACS, ...).
Besides MM/MD jobs, there are also some free energy, IR and thermodynamic integration simulations possible using programs such as AMBER. There are free interfaces such as VMD (http://www.ks.uiuc.edu/Research/vmd/) to visualize the results of our simulation although they will not submit your job and certainly not interactive. The AMBER tutorial do a good job of showing how to use the command line AMBER together with whatever free visualization tools are available (http://amber.scripps.edu/tutorials/). Ultimately, however, a GUI would greater broaden the user base of programs like AMBER (much as GUI have broadened the userbase for ab-initio programs) and make research quicker to perform.
The only interactive GUI that are available for MM/MD work are geared towards the pharmaceutical industry. They are, needless to say, VERY VERY expensive. Some reasonable examples are MOE (http://www.chemcomp.com/) and INSIGHT II (http://www.accelrys.com/products/insight/). I am sure that there are many others that I have omitted such as Schrodinger (http://www.schrodinger.com/).
In my opinion there is a rather large gap between the command line only MM/MD programs with little or no GUI to speed up the time needed to setup and submit jobs, and the full-blown pharmaceutical grade research tools. Whether you feel that WebMO is a good fit to fill this gap, is a question better left for the developers and the research public. I welcome any corrections or additional thoughts to my comments