AMBER Archive (2007)

Subject: Re: AMBER: PMF calculation

From: David A. Case (case_at_scripps.edu)
Date: Mon Jun 25 2007 - 19:47:25 CDT

On Mon, Jun 25, 2007, Jena M wrote:
>
> I am having some difficulties in creating the umbrella sampling input file.
> I want to calculate PMF of a ligand moving from 8 Å to 2 Å of a residue,
> keeping the protein residues within 5 angstroms of the ligand and of the
> residue frozen. I want to decrease the distance at 0.2 Å intervals. So, for
> instance for the first window do I have to set r2=r3=7.8?

Maybe you will also want a window at r2=r3=8.0 as well.

>
> #distance restrainst
> &rst iat=662,667, r1=0.0, r2=7.8, r3=7.8, r4=100.0, rk2 =100., rk3 = 100.,
> iresid=1,
> ATNAM(1)='SG',
> ATNAM(2)='C2',
> /

Your values of rk2,rk3 are pretty high: do a trial run, and see how much
variation you get in the values of the distance you are constraining. In
order to get good overlap with adjacent windows, you may well need to have a
somewhat smaller value. But this is something you can't tell until you try.

>
> As far as I understood we have two ways to prepare the input files for each
> window/simulation. One is to use the configuration of the previous window as
> input for the next window. The second one is to use always the same input
> file for each window. Is this correct? Which of the methods is better?

In any case, you have to equilibrate each window before collecting statistics
on it. The first option above is likely to led to the most rapid
equilibration, because the system will already be close to where you want to
be.

...good luck...dac

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