Forcefields


Generating the force field parameters


Different ways of doing the same task

Creating a parameter file can be a daunting task at times depending on which famous brand you are using. There are plenty of force fields available in the market and most of them are compatible with NAMD (but you need to specify which one you are using in the namd configuration file, more on that later). CHARMMS, OPLS, AMBER, etc are the commonly used ones. At times you might be required to copy the parameter values from the original publication, one by one, for all the atoms of the system. A diligent approach, something to scare away the amateurs trying to make a name in the field, and also to give a feel of "research" to the more serious ones. For our case we will be using OPLS-AA force field parameters. There is a web-application developed by the Jorgensen group (the ones who developed the OPLS parameters) which allows automatic generation of these parameters.

It is named LigParGen (short for Ligand Parameter Generator). Personally, I don't think the parameters are reliable. Always cross check. Automation cannot be a replacement for diligence, just an assistance. Also, it doesn't work for any complex or big molecule that you specify. So good luck finding your parameters here.

You might be suggested to look for these parameters from the original publication. This is a tedious work, and you need to depend on your instincts at times to decide which parameter to use. Also, keep adding references to it. Your publication will be in jeopardy even if you misspell a single value.


LigParGen

I don't know why, but the upload pdb file option rarely works in the site. Thanks to Leela Dodda? Here are a few steps to help you out with the parameter file generation.



Python?

We have the parameter file, but it has atoms names in the form of C800 or C801, etc. We need to specify the atom names as per the atom types we had used in the psf file generation step (rings a bell?). What you need to do? Open the pdb file in notepad++, it has different names (C01, maybe). We need to connect these with C800 and then edit that to our atomtype names. A bit of utterly futile task, which could have been avoided by the webapp developers (Leela dodda?), but it is what it is. You can write a python script to automate this task. I'll share if I get time. Use VMD to figure out which atom corresponds to which atomname, the .rtf file connects those pdb file names to the ones given in .prm file. COnfusing? Thank Leela dodda?


Previous Page          Main Page          Next Page