Date: Wed, 31 Jan 2001 11:28:43 -0800
From: David Case
Subject: Re: Belly option in sander

On Wed, Jan 31, 2001, Florian Barth wrote:

>
> I would like to run a md simulation in sander using the belly option to
> constrain the protein backbone. This is my md.in:
>
> #50ps simulated annealing with temp. ramp and Belly
> &cntrl
> nstlim=25000, dt=0.002, ntt=1, ibelly=1, igb=3, ntb=0, ....
> &end

> "When igb>0, the moving part must be at the start of the molecule.
> This does not seem to be the case here."
>
> Does anybody know, how the md.in or the topology files must be altered
> to run above simulation in sander. A colleague pointed out to me,
> that it is possible with
> sander_classic to run belly simulations but I would like to stick to
> the "new" sander.

I'm sending this response to the whole list, since this point may be
confusing to many people.

If you are simulating a non-periodic system with ordinary Coulomb or
distance-dependent dielectric (as above: ntb=0,igb=3), in almost all
circumstances you should sander_classic rather than sander. Sander_classic
is both more flexible and faster for these sorts of calculations.

People should not assume that "new" sander is somehow better that
sander_classic. Eventually we will be able to remerge the codes, but for
now, as a rule of thumb:

==============================================================================
For periodic simulations with PME, and for non-periodic simulations with
generalized Born: use sander.

For all other non-periodic simulations, and to use OWFEG, use
sander_classic.

If you really want to run a periodic simulation without PME (i.e. with
cutoffs instead), use sander_classic; (in my personal view, however,
this sort of simulations is not generally to be recommended.)
==============================================================================

It would clearly be better if we had one program that was "good" at
everything, and apologize for any confusion that the current situation
leads to.

...hope this helps...dac