AMBER Archive (2006)

Subject: RE: AMBER: Question regarding lastrst and taup

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Wed Oct 25 2006 - 10:46:52 CDT


Dear Priti,

> I have a basic question regarding lastrst and taup values. I
> have done two
> simulations on the same protein at a high temperature (600K) for only
> 500ps. The box size of the two simulations are different (8 and 12
> Angstrom respectively) and all the other initial conditions
> were same. But
> the simulations had stopped saying that I should increase
> lastrst. So in
> the first simulation I increased the lastrst value to 5000000
> and changed
> taup value to 2.0 whereas in the other simulation I kept the
> taup value
> 1.5 and had to change the lastrst value to 15000000 as
> suggested by the
> error. Now when I check the RMSd values, in the first simulation it
> reaches around 8 Angstrom and in the second one its around 15 Angstrom
> after 500ps of simulation ! Is this kind of different result
> expected with
> just different box size and different values of taup and lastrst ? My
> protein after 500ps of first simulation is quite intact whereas in the
> second one its very much distortd !

Unless there is something wrong with your simulation you should never need
to change the value of lastrst. There is a reason it does not appear in the
manual.

There is also no relationship between lastrst and taup. The first controls
memory allocation while the second controls the barostat relaxation time.
Unless you have a specific reason for wanting to change the pressure
relaxation I would leave taup at the default value. Instability in the
trajectory often comes from other places 'before' it comes from the pressure
coupling.

It would be useful to see your input files and the protocols you used. If
you hot started at such a high temperature you are almost certainly likely
to have an unstable system. At 600 K you may also need to reduce the
timestep from say 2fs (with shake) to 1.5fs or so.

Also, are your boundary conditions correct? NTB=2 and NTP=1?

You also note that the calculation of pressure is very inaccurate at low
temperatures and so the barostat can over correct leading to instability.
You should always start your heating from 0K with constant volume (NTB=1)
and then after say 20ps or so when you are above 100K you can switch to
constant pressure (NTB=2).

It would also be useful to see what you set the cutoff value to. Typically
lastrst errors come from setting a massive cutoff. If you are running a
periodic boundary simulation with PME then you only need a cutoff of 8
angstroms.

I am also not sure what you mean by box size of 8A - this seems incredibly
small. Do you mean this is the buffer size you used in Leap when solvating
the system? If this the case then your box size will be 16A + the diameter
of your molecule in X Y and Z directions. You can check the output file
where the box size should be printed.

With regards to the RMSD both values seem very high and suggest a problem
with the trajectory. This assumes of course that you are measuring the RMSD
of the protein and not of the whole system. Since water molecules are not
reimaged during a PME simulation your RMSD will just keep increasing as you
run a simulation. Hence the rmsd you are interested in is the solute RMSD.

All the best
Ross

/\
\/
|\oss Walker

| HPC Consultant and Staff Scientist |
| 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.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu