Re: Is very slow heating "bad"?

From: Vincent Kraeutler (vincent_at_kraeutler.net)
Date: Sat Aug 04 2007 - 11:06:11 CDT

Hi,

Since you ask for "any thoughts", here are some of mine ;-)

First of all, I suppose it really depends on what you want to see in the
unfolding process.

Personally, I'm not such a big fan of "heating" since it has little to
do (but is often mixed up) with heating up a sample in a wet lab. How
would you reproduce a heating rate of 10^12 K / s (or even 10^9 K / s)
in the real world? ;-)

I'm not an unfolding expert, but I have experience with the related
problem of equilibration. Here, I'd rather go for gradually reduced
position restraints starting right at your target temperature. The
advantages are three-fold (and I think they carry over to your unfolding
example):

-- you simulate under known thermodynamic boundary conditions;

-- you can directly read out the forces required to maintain the current
structure, and therefore have a feeling of whether you're going in the
right direction;

-- you can step in at any time if the system goes haywire.

What you should use for the relaxation time scales depends on what you
intend to simulate. But the following relaxation times are (IMHO) a
reasonable start:

solvent shell: 0.5-2 ns
sidechains: 5-50 ns
ion distributions: 10 - 100 ns

At the risk of straying very far from your original question, I think an
unfolding protocol could look like this:

simulate with restrained solute until the solvent shell is relaxed (~ 2 ns).
simulate with restrained backbone (~ 5 ns).
simulate unconstrained system (until unfolding criterion is fulfilled).

You can then immediately observe the effect of temperature-induced
desolvation (first step), the corresponding change in sidechain
interactions (second step, you would expect stronger salt-bridges and
weaker hydrophobic contacts), and finally, the unfolding of the protein
structure (if at all, unfolding can take a long time!).

Cheers,
v.

Arun Krishnan wrote:
> Hi All,
>
> When heating, I found that people have used very slow heating (heating
> linearly to 300ps) to bring up
> the temperature from 0K - 300K)
>
> reassignFreq 1
> reassignIncr 0.001
> reassignHold 300
> numsteps 300000
>
> timestep = 1fs
>
> So basically, increases the temperature by 0.001 degK every step 1degK/ps
>
> Is that too slow and will it affect my simulation later on? I realize
> that increasing the heat too fast can lead to a massive input of
> kinetic energy
> in too short a time and that can cause the system to go haywire.. but
> what about too slow heating? Since I am trying to unfold the protein
> I want to make sure that the heating is slow and nice. Any thoughts?
>
> Regards,
>
> Arun
>
> --
> ***********************************************
> Arun Krishnan, Ph.D,
> Assistant Professor,
> Institute for Advanced Biosciences,
> Keio University,
> Center Building,
> Tsuruoka, Yamagata 997-0035
> Japan
> Phone: +81 (0)235-29-0824
> Email: krishnan_at_ttck.keio.ac.jp <mailto:krishnan_at_ttck.keio.ac.jp>
> URL: http://www.iab.keio.ac.jp/~krishnan
> <http://www.iab.keio.ac.jp/%7Ekrishnan>
> **********************************************


This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:46:44 CST