Re: Fwd: Coarse Graining Problem

From: Anton Arkhipov (anton_at_ks.uiuc.edu)
Date: Wed Feb 24 2010 - 11:20:08 CST

Hi,

You can check the description of CG Tools in VMD here:

http://www.ks.uiuc.edu/Research/vmd/plugins/cgtools/

At the very bottom of that web page, there is a description of how to
create arbitrary CG bead definitions, with an example for water. If
you follow that example, you can create any CG model for amino acids,
including 4 or 6 beads. You'll then also need to create topology and
parameter files for such a model. If potential forms are compatible
with what NAMD can represent (e.g., same potential forms as used in
all-atom CHARMM force field), then you should be able to simulate the
resulting CG systems with NAMD.

Since residue-based CG model is more detailed than shape-based CG,
it's also supposedly more accurate. Shape-based CG permits you to
study larger systems on longer time scales. There was a number of
rather successful studies of protein/lipid assembly using the residue-
based CG, for example see nanodisc/HDL papers cited here:

http://www.ks.uiuc.edu/Research/CG/

Ultimately, though, CG methods are not very accurate, and particularly
it's hard to trust them if you want to study something like amyloid
aggregation, where fine details may matter a lot. So, one needs to be
careful and test one's models thoroughly before trusting them.

Anton.

On Feb 24, 2010, at 3:59 AM, Aditya Ranganathan wrote:

>
>
> ---------- Forwarded message ----------
> From: Aditya Ranganathan <aditya.sia_at_gmail.com>
> Date: Wed, Feb 24, 2010 at 4:07 PM
> Subject: Coarse Graining Problem
> To: majordomo_at_ks.uiuc.edu, vmd-l_at_ks.uiuc.edu
>
>
> subscribe vmd-l
>
>
> Respected Sir,
> I`ve been looking at various coarse graining methods available.
> However, the CG Tools Residue Based Coarse Graining technique uses a
> 2 bead model of coarse graining (if I`m not wrong). However, from
> waht I gather, 2 bead models are`nt accurate enough representations
> to study more complex events like aggregation. I came across
> literature on the internet where there have been mention of four or
> even six bead models.
>
>
> Can you clarify whether a 4 bead model can be implemented and used
> in VMD/NAMD or by any other tool?
> Also, how would you rate the Shape Based Coarse graining methods
> compared to the REsidue Based Method in terms of accuracy in case of
> a protein aggregation/assembly study.
>
>
>
> Thanks
> Srivastav Ranganathan
> Junior Research Assistant,
> Department of Biosciences and Bioengineering,
> Indian Institute of Technology, Mumbai, India
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:22:45 CST