From: Gabriel Jara (gabriel.fcq.unc.ar_at_gmail.com)
Date: Fri Jul 01 2016 - 16:13:32 CDT
Hi Gumbart,
Thank you for the answer. Now, I understand that the convergence is slow,
so I focued in the stratification strategy in order to improve the results,
do u have any tips to this issue?
I looked at the tutorial, but I am not sure if it is what I looking for,
because I am intereted in a PMF profile more than of the total variation of
free energy.
Thank again.
Gabriel
On Thu, Jun 30, 2016 at 5:51 PM, JC Gumbart <gumbart_at_physics.gatech.edu>
wrote:
> Hi Gabriel,
>
> I would suggest taking a look at the following tutorial and associated
> papers:
> http://www.ks.uiuc.edu/Training/Tutorials/namd/PLB/tutorial-protein-ligand.pdf
>
> Calculating a binding free energy can be very challenging! Particularly
> for a flexible ligand like ATP. And, to be honest, ABF has a great
> tendency to produce values that are far too high initially (due to
> non-equilibrium effects), then requiring potentially hundreds of ns to come
> back down.
>
> Best,
> JC
>
> On Jun 30, 2016, at 2:25 PM, Gabriel Jara <gabriel.fcq.unc.ar_at_gmail.com>
> wrote:
>
> Dear all,
>
> I am interested in studying the process and free energy of binding of ATP
> to the active site and coordinated to a magnesium. The strategy I thought
> was to start from the protein:ligand complex and to perform ABF
> simulations in order to calculate the PMF of unbinding and in this way to
> describe the mechanism of binding and its free energy.
>
> The problem is that I obtained huge free-energy values, ~150 kcal/mol with
> simulation of 80 ns. I tested different reaction coordinates, but the
> results are similar. At the moment, I am testing a open-conformation of the
> protein, in order to know the conformational effect.
>
> The reaction coordinate selected is a distance between the center of mass
> of the whole ligand or the nitrogenenous base or the phosphate, and a
> non-flexible loop in the active site. In addition, the magnesium is
> released from the active site along with ATP, only if I explicitly add it
> in the reaction coordinate.
>
> I have little experience about ABF, and I would like to know tips to
> overcome this issue.
>
> Following attached the collective variable file
>
>
> Colvarstrajfrequency 1000
> Colvarsrestartfrequency 1000
> colvar {
> name COMDistance
> width 0.1
> lowerboundary 7.00
> upperboundary 20.00
> lowerwallconstant 10.0
> upperwallconstant 10.0
> distance {
> group1 {
> atomnumbers { 22 23 24 25 26 27 28 29 30 31 32 33 34 35 }
> }
> group2 {
> atomnumbers { 5461 5462 5463 5464 5471 5472 5473 5474 5475 5476 5492 5493
> 5494 5495 5496 5497 5504 5505 }
> }
> oneSiteSystemForce yes
> }
> }
>
> abf {
> colvars COMDistance
> fullSamples 1000
> hidejacobian yes
> }
>
> Other question, is better to separate the whole reaction coordinated in
> several windows, similar to the next work?
>
> dx.doi.org/10.1021/ci400188q J.Chem.Inf.Model.2013, 53, 2376−2389
>
> Thanks in advance.
>
> Gabriel
>
>
> --
> Dr. Gabriel E. Jara
> Instituto de Química / UNICAMP
> Rua Josué de Castro s/n
> Cidade Universitária "Zeferino Vaz", Barão Geraldo
> 13083-861 Campinas, São Paulo, Brasil
> ------
>
>
>
-- Dr. Gabriel E. Jara Instituto de Química / UNICAMP Rua Josué de Castro s/n Cidade Universitária "Zeferino Vaz", Barão Geraldo 13083-861 Campinas, São Paulo, Brasil ------
This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:34 CST