> I have some questions about SANDER and PMEMD. I tried to run GB simulation
by
> SANDER and also tried the TIP3P water simulation by PMEMD.
> The system
> is a 100 residues protein(charge +9) and the water number is ~8000.
> The GB/SANDER is
> much slower than the TIP3P/PMEMD. It surprised me because I
> always think that
> implicit solvent is much faster than explicit one. Is it due
> to the different
> between PMEMD and SANDER? If it is, is there any way to run
> the GB simulation
> with higher speed?
You are really trying to compare apples and oranges here. The first thing
you should do is use pmemd for both the PME and GB simulation. PMEMD v9.0
supports GB simulations.
Next you should look at the theory involved in PME simulations and in GB
simulations. From this you will quickly see that for Generalized Born
simulations the functional form of the equation is much more complicated and
thus per atom atom interaction more work has to be done. Secondly you have
to do more of these interactions due to the need to use much larger cutoffs
for GB simulations than you do for PME simulations. Although you get a lot
of this back in that the total number of interactions is typically less due
to there being less atoms and also that you don't have to do the reciprocal
calculation.
The issue is that the two calculations types scale differently with the
number of atoms. For small numbers of atoms (a few thousand) implicit
solvent GB will generally be faster than the same system solvated. However,
as you go to larger numbers of atoms for GB the calculation time will
quickly exceed the equivalent solvated system time.
You should think carefully about using implicit verses explicit solvent. For
a medium sized protein the difference GB can actually be slower than a
solvated system so you need to think about why you want to do an implicit
solvent simulation. You shouldn't just do it because you want to save time.
Instead implicit solvent simulations should be run where it is beneficial to
do so. E.g. if you want to calculate free energies of solvation etc. There
is also the argument that you may need less equilibration time with an
implicit solvent calculation which can benefit you overall. All in all you
need to think carefully about what you want to learn from the simulation.
> The other thing is that PMEMD seems not support the ibelly
> simulation,
> which can fix specific part of protein. Is there any way to do it?
Do you actually need to use Ibelly? You can use restraints to fix a part of
the system and pmemd supports this.
All the best
Ross
/\
\/
|\oss Walker
| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross.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.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Oct 08 2006 - 06:07:08 PDT
