[gmx-users] Re: Umbrella sampling with temperature and pressure coupling method problem (Justin A. Lemkul)

[Justin A. Lemkul]

DeChang Li wrote:
> 
> 
> 
>     Message: 6
>     Date: Tue, 19 Oct 2010 09:30:47 -0400
>     From: "Justin A. Lemkul" <jalemkul at vt.edu <mailto:jalemkul at vt.edu>>
>     Subject: Re: [gmx-users] Umbrella sampling with temperature and
>            pressure        coupling method problem
>     To: Discussion list for GROMACS users <gmx-users at gromacs.org
>     <mailto:gmx-users at gromacs.org>>
>     Message-ID: <4CBD9D87.6080404 at vt.edu <mailto:4CBD9D87.6080404 at vt.edu>>
>     Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> 
> 
>     chris.neale at utoronto.ca <mailto:chris.neale at utoronto.ca> wrote:
>      > Use Langevin dynamics (the sd integrator) to control the temperature.
>      > You are sure to get the correct ensemble that way and if you are
>     doing
>      > US then you can not extract dynamics anyway.
>      >
>      > Hopefully somebody else can address the pressure coupling for
>     you, but
>      > probably you need to provide more information to get a useful
>     answer there.
>      >
> 
>     The Berendsen barostat suffers from the same limitations as the
>     thermostat - the
>     pressure distribution does not produce a true NPT ensemble.
> 
> 
>    Whether the correct canonical ensemble (NPT or NVT) is important (or 
> indispensable) for umbrella sampling? If I used the weak coupling method 
> (Berendsen) to do the simulations, can I extract the PMF from the 
> umbrella sampling simulations?
>  

I would think that a proper statistical mechanical ensemble would be considered 
indispensable for any simulation.  In my mind, there is no real reason to use 
less accurate methods when collecting data.  If your goal is a comparison of 
methods, or consistency with other results, sure, then you may have a reason to 
use algorithms that may be less than optimal.  For any sensitive thermodynamic 
study, I would strongly argue that your potential energy surface needs to be 
rigorously correct.

The point is this.  You have to defend your choices to a skeptical audience 
(reviewers).  This is one question that might be asked, and really should be. 
Methods should be scrutinized.  So I would ask you this: why perform your 
simulations with algorithms that are not as accurate as others, when those 
better algorithms are accessible to you and do not harm performance in any 
demonstrable way?

-Justin

> 
> 
>     -Justin
> 
>      > -- original message --
>      >
>      > Dear all,
>      >
>      >     I want to use umbrella sampling to calculate the PMF of the
>      > conformational transition of a protein. What temperature coupling
>     method
>      > and
>      > pressure coupling method should I use? Berendsen temperature
>     coupling or
>      > Nose-Hoover temperature coupling? Or each one is OK?
>      >
>      >
> 
>     --
>     ========================================
> 
>     Justin A. Lemkul
>     Ph.D. Candidate
>     ICTAS Doctoral Scholar
>     MILES-IGERT Trainee
>     Department of Biochemistry
>     Virginia Tech
>     Blacksburg, VA
>     jalemkul[at]vt.edu <http://vt.edu> | (540) 231-9080
>     http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
> 
>     ========================================
> 
> 

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================