[gmx-users] gmx-users

[Mark Abraham]

Luis Javier Alvarez Noguera wrote:
> Dear colleagues:
> We have the cartesian coordinates of a micelle in xyz format and haven't been
> able to build up the .top corresponding file with gromacs package program
> x2top. There is a confusion on how to feed this program. Is there a document
> or an example where we can clarify thigs such as:
> 1. When building the micelle by hand, what force fields are needed?
> 2. When is it advaisable to use united  atoms such as CH3, CH2,  etc. and when
>  one should  use  single atoms?
> 3. How are atoms names related to the force field?

You seem to have the approach a bit backwards. The choice of force field 
is the one which drives everything else, and the basis for that choice 
should be what you are hoping to achieve with the simulation. That means 
reading literature to see what other people have done, and adapting 
their approach to your conditions. A secondary consideration is whether 
you have access to suitable tools to generate a suitable physical 
description of your starting configuration (for GROMACS, generating the 
.top file).

So that takes care of question 1. Question 2 is related to it, but 
united-atom force fields can use longer integration timesteps, and are 
cheaper to compute per-timestep too. The atom names are chosen by the 
force field designers, and you may need to change your input structure's 
atom names so that an automatic-topology-generation tool can do the 
right job.

Mark