AMBER Archive (2009)

Subject: RE: [AMBER] Tutorial A3

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Mon Jul 20 2009 - 14:03:44 CDT

Hi Bill,

> In tutorial A3 (MM-PBSA calcualtion), in section 1 during heating
> stage, you used two different temperaturre contorls, is this normal?

Firstly, please realize that these are mostly arbitrary choices. The purpose
of initial minimization and heating is simply to get the system from its
initial, strained crystal structure, to something that has equilibrated out
the biases of the starting configuration and is stable and at the correct
temperature to start 'measuring'. In principle you could do this any way you
pleased as long as you could ultimately show that your heating /
equilibration algorithm did not affect your results. If it did, and you
weren't directly trying to measure this, then you would have serious
problems. Thus one typically chooses a minimization, heating and
equilibration algorithm to be as gentle as possible. Obviously being too
gentle means it takes a lot of time and effort. Thus there are tradeoffs to
be made. Typically these are problem specific. Often in implicit solvent one
can be more aggressive than explicit solvent. Large systems tend to be more
stable than smaller ones etc. Short answer is do not read too much meaning
into the specific algorithms used in the tutorials.

>
 is
> it enough to define only temperature variation by using :
> &wt TYPE='TEMP0', istep1=0, istep2=25000,
>   value1=0.1, value2=300.0, /
>  &wt TYPE='END' /

I do not understand what you mean here. Here is the input file from the
tutorial:

heat ras-raf
 &cntrl
  imin=0,irest=0,ntx=1,
  nstlim=25000,dt=0.002,
  ntc=2,ntf=2,
  cut=8.0, ntb=1,
  ntpr=500, ntwx=500,
  ntt=3, gamma_ln=2.0,
  tempi=0.0, temp0=300.0,
  ntr=1, restraintmask=':1-242',
  restraint_wt=2.0,
  nmropt=1
 /
 &wt TYPE='TEMP0', istep1=0, istep2=25000,
  value1=0.1, value2=300.0, /
 &wt TYPE='END' /

This looks to be the same as what you are suggesting.

> Also, in Minimization stage, you minimized solvent only with a
> constrain on your protein-ligand? why you didn't minimzed protein-
> ligand complex also?

We probably should have but the restraints were weak enough that any large
clashes would get minimized our anyway.

> Why in the production step, he used constant pressure, not constant
> volume?

I think it would make little difference to the results to use either. Try
it. Check the density is good and rerun with NVT or NVE - obviously making
sure you run production for a decent length of time, not the abbreviated
amount used in the tutorial.

> I would appreciate if you can direct me to one reference
> dealing with the difference between using constant presssure or volume
> in the production phase.

This is really not a simple question that can be dealt with easily. I know
of no references that really deal with this, maybe someone else on the list
can suggest something. The short answer is that things are really not so
black and white. This is research, there is more than one way to do things
and no simple RIGHT answer. Often people use NVT or NVE because it is,
perhaps, one less approximation, I.e. no barostat present. You can also get
performance advantages (especially in parallel) for using NVT / NVE over
NPT. In this case though the property being measured should not be dependent
on pressure since we are using the MD to generate snapshots for MMPBSA so in
reality the choice is largely arbitrary.

All the best
Ross

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber