AMBER Archive (2008)

Subject: Re: AMBER: Heating up

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Fri Nov 28 2008 - 09:57:49 CST

Thanks. Actually, I was inaccurate: I used to raise gradually
(nstlim=25000, Langevin) the temp from 0.1K to 300K. This is also what
I did now for the protein in a membrane+water. RMSD OK. But, you made
the point quite clear. I was lucky.

A basic difficulty for me is to imagine to start with a system at 100K
when it was 0K (at least in terms of QM)

francesco

On Fri, Nov 28, 2008 at 1:30 PM, Carlos Simmerling
<carlos.simmerling_at_gmail.com> wrote:
> I advise against heating from zero in any case. the details depend a
> bit on the thermostat type, etc, but I think it can still be dangerous
> to ramp the temperature too much. an extreme example is with a
> Berendsen thermostat- if part of the system is a bit high in energy
> when you start, even if it is in a local minimum, it can result in
> very hot sections of the system. for example, if you start at 10K and
> a side chain was strained and picks up just a little extra kinetic
> energy , say enough to put it at 50K, then when you scale the bath to
> 300K, the 50K part gets scaled to 1500K. If you start at 100K and part
> is 50K hotter, then it's not nearly as bad a problem.
>
> ramping the temperature by a large factor in a short run requires a
> really good thermostat and good coupling between various parts of the
> simulated system (which can be worse in GB, which may be why you were
> advised to be especially careful there).
>
>
>
> On Thu, Nov 27, 2008 at 1:29 PM, Francesco Pietra <chiendarret_at_gmail.com> wrote:
>> For GB, I was recently advised to start heating up the system to 300K
>> from about 100K, not from 0K. Now that I am back to explicit
>> environment for a protein complex (25,000 atoms) partly in a lipidic
>> membrane, and the whole in TIP3BOX (68,000), where I was used to
>>
>> nmropt=1
>> /
>> &wt type='TEMP0', istep1=0,istep2=35000,
>> value1=0.0, value2=300.0
>> /
>> &wt type='END'
>> /
>>
>> does a similar warning as for GB applies here too?
>>
>> Thanks
>> francesco pietra
>> -----------------------------------------------------------------------
>> The AMBER Mail Reflector
>> To post, send mail to amber_at_scripps.edu
>> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
>> to majordomo_at_scripps.edu
>>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu