ABF: example works now; more questions.

From: Jan Saam (saam_at_charite.de)
Date: Tue May 30 2006 - 08:57:15 CDT


thanks for the clarification, I got the example working now!

May I bug you with one more question?
I'm studying the diffusion of oxygen along a channel in a protein. The
channel is not straight but follows despite some kinks roughly along a
certain direction from the active center to the protein surface. There
are also some "bays" or dead end roads along the tunnel that can be
explored by oxygen.

Thus, most of the time the gas molecule will not travel exactly along
the defined reaction coordinate.
Does that mean the following?:

a) If distance active center -- O2 is the RC:
Oxygen could be temorarily trapped in a dead end road, its hammering
against the wall and the adapdive biasing force is cranked up until
oxygen breaks through that wall? And even when it's travelling back,
finding the right way, the time spent in the side channel will distort
my PMF?

b) If abscissa along main direction is used as RC:
When I'm defining a direction of the RC there will always be a certain
angle (say between 0 and 50 deg.) between the real channel V and the RC
due to the crookedness of the channel. I.e. the adaptive biasing force
is pushing with that angle against the channel wall, which will also
distort my PMF. Correct?

c) If I'm partitioning the channel into several shorter, almost straight
portions, I can stitch the PMF together and probably improve on these
problems. But I'll have to know the RC very well before and it is not
easy to define the directions in a system where everything moves.

Could you please comment on that, especially items a) and b)? I'm sure
you have experience with that.


PS. I'm not planning to use rgidbonds or any other additional
constraints whatsoever.

Chris Chipot wrote:

> Jan:
>> Dear Chris,
>> thanks for your quick answer!
>>> a) what are the atoms corresponding to abf1 4 and abf2 117 ? Keep
>>> in mind that the RC should be decoupled from constrained DOFs.
>> atoms 4 and 117 are the C-alpha atoms of the first and the last
>> residue. (In contrast to what I erroneously stated in my previous
>> mail I used dodeca-alanine).
> If your peptide is shaken/rattled (rigidBonds all), then your
> free energy calculation is likely to be erroneous. I would like
> to emphasize once again that the RC ought to be fully decoupled
> from frozen degrees of freedom. If rigidBonds is set to all, to
> allow longer time steps to be used, note that the Calpha...Calpha
> RC is coupled to shaken/rattled Calpha-Halpha bonds.
> There are three routes you can follow to get the correct answer:
> i) turn rigidBonds to none, in which case you can keep the
> Calpha...Calpha RC, but ought to pay attention to the time
> step you are using.
> ii) use the distance separating the first and the last carbonyl
> carbon atoms as your RC. These atoms are not coupled to any
> frozen DOF.
> iii) keep the Calpha...Calpha as your RC and rigidBonds to all. Turn,
> however, to another type of ABF coordinate: distance-com, and
> include the Halpha in the sets of atoms for abf1 and abf2.
> Literally, your RC will be the distance separating the COMs
> of the first and the last Calpha-Halpha bonds, which is close
> enough to a true Calpha...Calpha distance.
>>> b) note that the free energy minimum of the alpha-helix emerges
>>> around 14 angstroms. In your input file, xiMin is 18.
>> I just started with the folded helix where the distance between the
>> atoms is about 18 A (for dodeca-alanine; meanwhile I tried the same
>> with deca-alanine and got the same bad sampling). The goal was to
>> unfold the helix starting from the perfectly olded state.
> See answer to point a). I can guarantee that your sampling will improve
> instantly if you respect the above basic requirements for running ABF
> free energy calculations. In a matter of 2 ns, you should get a very
> reasonable estimate of the stretching free energy profile.
>>> c) the role of dxi is important and should be chosen as a function
>>> of how rapidly the free energy is changing. 0.5 angstroms is in
>>> the case of deca-alanine probably too large. 0.1-0.2 angstroms
>>> is typically what we used. dSmooth should set accordingly.
>> I tried values between 0.5 and 0.02. The potential form changes but
>> the sampling is equally bad in all cases.
> See answer to point a). Pick either i), ii) or iii) and switch to a
> 0.1-0.2-angstrom dxi. Update dSmooth accordingly, ca. 0-0.2. In the
> gas phase, the threshold for applying the average force, FullSamples,
> can be set to a rather low value, viz. typically 100 samples.
> Chris Chipot
> _______________________________________________________________________
> Chris Chipot, Ph.D.
> Equipe de dynamique des assemblages membranaires
> Unité mixte de recherche CNRS/UHP No 7565
> Université Henri Poincaré - Nancy 1 Phone: (33) 3-83-68-40-97
> B.P. 239 Fax: (33) 3-83-68-43-87
> 54506 Vand½uvre-lès-Nancy Cedex
> E-mail: Christophe.Chipot_at_edam.uhp-nancy.fr
> http://www.edam.uhp-nancy.fr
> Science without management is worse than management without science
> N. G. van Kampen
> _______________________________________________________________________

Jan Saam
Institute of Biochemistry
Charite Berlin
Monbijoustr. 2
10117 Berlin
+49 30 450-528-446

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:42:06 CST