From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Sun Nov 25 2012 - 00:46:48 CST
The sign difference exists for an obvious reason - because one direction is creation and one is annihilation. In the plot, the reverse direction was just negated in order to show the two curves exhibit no hysteresis. In Figure 10, I think only one direction is plotted, because here, there is a notable difference between forward and backward directions using the supplied files, suggesting that they aren't fully converged. Try running the sims yourself for much longer to see if this discrepancy between directions is reduced.
On Nov 24, 2012, at 9:50 PM, dbaogen wrote:
> Dear Dr. Chipot,
>
> Thank you for your detailed instructions! I make a mistake of description in the previous email. In fact, I used the fep.tcl to carry out FEP calculation and obtained the free energy change from 0 to 1 and confused energy and free energy. Now I understand this after your instructions. Thanks!
>
> In addition, I have some problems to reproduce the results shown in Tutorial FEP (http://www.ks.uiuc.edu/Training/Tutorials/#namd ). For example, charging of a naked Lennard-Jones particle into a sodium ion, I extracted the values of free energy change using deltaA.awk script for forward and backward fepout download from http://www.ks.uiuc.edu/Training/Tutorials/#namd . And free energy profile obtained is not the same as the Figure 5 in Tutorial-FEP.pdf file. In Figure 5, forward and backward parts are overlapped. In fact, the values of free energy change for the forward and backward have a sign difference. How could I obtain the same free-energy change file as Figure 5. The same condition occurred in the "Binding of a potassium ion to 18-crown-6" part. Using the downloaded *.fepout files, I could not reproduce the same free energy profile as the Figure 10. Whether do some tricky techniques to deal with the free energy change data? Thank you in advance!
>
>
> Best wishes
>
> Duan Baogen
>
>
> 发件人： Chris Chipot
> 发送时间： 2012-11-24 23:54
> 收件人： dbaogen
> 抄送： namd-l
> 主题： Re: namd-l: question about relative binding free energy calculation method using FEP
> Dear Baogen,
>
>> I want to calculate the relative free energy between RAN polymerase and ATP or GTP using FEP method implemented in NAMD software. Firstly, the mutation of ATP to GTP in water solution is calculated. After 100 ps equilibrium of the system, forward and backward FEP analysis are carried out. The energy values are close when lambda value is larger than 0.5. But there exists deviation when lambda value is less than 0.5. what will result in this condition?
>
> Who knows! What can be predicted, however, from the values assigned
> to alchVdwLambdaEnd and alchElecLambdaStart is a discontinuity
> of ∆G at λ = 0.5, when only the van der Waals term of the intermolecular interactions subsist for the perturbed portion of your system. This aspect
> of free-energy calculations is actually discussed at length not only in the
> user's guide, but also in the dedicated FEP tutorials.
>
>> The FEP parameter setting is as follow:
>>
>> alchVdwLambdaEnd 1.0
>> alchElecLambdaStart 0.5
>> alchVdWShiftCoeff 4.0
>> alchDecouple on
>>
>> Another question is how to obtain the energy value when lambda is equal to 0 or 1 corresponding to forward or backward analysis?
>
> Your question is not particularly clear. There seems to be a confusion
> between energy and free energy. Since you are running FEP calculations,
> what you are measuring is not a free-energy change at λ, but a free-energy
> change between λ and λ + δλ. Conventionally, in a forward free-energy
> calculation, ∆G is 0 at λ = 0. Likewise, in a backward transformation,
> ∆G is 0 at λ = 1.
>
>> If the initial value of lambda is 0, and increment is 0.05. when the calculation is completed, the output *.fepout file will give the energy value from 0.5 to 1.
>
> No. Unlike the starting point of your simulation is set to be 0.5 and its endpoint
> to be 1. I can assure you that if you use fep.tcl and write
>
> runFEP 0.0 1.0 0.05 $nSteps
>
> your output file will, indeed, contain all the free-energy changes from 0 to 1,
> i.e., 0 to 0.05, 0.05 to 0.1, etc... 0.95 to 1.
>
>> Does the energy value correspond to lambda(0) state when I set the FEP parameters alchLambda (0 )and alchLambda2 (0) simultaneously?
>
> If you set concomitantly alchLambda and alchLambda2 to 0, you will
> compute a free-energy change between λ = 0 and λ = 0 -- and find out
> that ∆G is 0. Computer time well spent.
>
>> Would you like to help me with the problems?
>
> If you do not find a nil ∆G between λ = 0 and λ = 0, please, write to us.
>
>> Thank you for your great help!
>> Best wishes
>>
>> Duan Baogen
>
> Chris Chipot
>
> --
> _______________________________________________________________________
>
> Chris Chipot, Ph.D.
> Theoretical and Computational Biophysics Group
> Beckman Institute
> University of Illinois at Urbana-Champaign
> 405 North Mathews Phone: (217) 244-5711
> Urbana, Illinois 61801 Fax: (217) 244-6078
>
> E-mail: chipot_at_ks.uiuc.edu
> Christophe.Chipot_at_edam.uhp-nancy.fr
> Web: http://www.ks.uiuc.edu/~chipot
> http://www.edam.uhp-nancy.fr
>
> The light shines in the darkness, and the darkness has not overcome it.
> John 1:5.
> _______________________________________________________________________
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:47 CST