Dear Samuel,
Since I implemented the TI decomp I can answer your question regarding this part:
The difference you obtain is actually no error.
The reason for the difference is just that the average given in the CHECK DECOMP section is the average over the dVdl-values that have been written out every NTPR steps which is also true for the different components of dVdl, whereas the dVdl value and its components written in the DV/DL, AVERAGES OVER <NSTLIM> STEPS steps is as its title says the average over the values of all NSTLIM steps.
Cheers!
Hannes
-------- Original-Nachricht --------
> Datum: Fri, 19 Feb 2010 10:10:45 +0100
> Von: Samuel Genheden <samuel.genheden.gmail.com>
> An: AMBER Mailing List <amber.ambermd.org>
> Betreff: [AMBER] TI calculation of absolute dG with restraint and energy decomposition
> Hello
>
> I've previously posted questions about my efforts to calculate
> absolute dG of benzene to T4-lysozyme using TI in Amber 10. I will
> follow the work by Roux, as cited in the Amber manual (JCTC 2 p1255).
> This means that I want to add a restraint to my benzene molecule, but
> I would also like to do an energy decomposition. The restraint energy
> should not be included in the dv/dl energies so I set dvdl_norest to
> 1. Now I have run a few lambdas on the van der Waals-part, i.e.
> perturbing an uncharged benzene to nothing but I'm confused about the
> output of sander. Here is a part of the output for one of the lambdas
>
> DV/DL, AVERAGES OVER 100000 STEPS
>
>
> NSTEP = 100000 TIME(PS) = 320.000 TEMP(K) = -0.17 PRESS =
> 0.0
> Etot = -8.9561 EKtot = -8.9561 EPtot =
> -22.2126
> BOND = 0.0000 ANGLE = 0.0000 DIHED =
> 0.0000
> 1-4 NB = 0.0000 1-4 EEL = 0.0000 VDWAALS =
> 8.8226
> EELEC = 0.0000 EHBOND = 0.0000 RESTRAINT =
> -31.0353
> EAMBER (non-restraint) = 8.8226
> DV/DL = -53.2479
> Ewald error estimate: 0.3153E-14
>
> ------------------------------------------------------------------------------
> CHECK DECOMP - TOTAL ENERGIES (w/ REST)
>
> INTERNAL= 0.0000
> VDWAALS = 8.8135 EEL = 0.000
>
> In a usual TI run, without restraint the EPTot is equal to DVDL, as
> EPTot = V1-V0. But EPtot is also, in this case, VDWAALS+RESTRAINT as
> it should be. But then I see that DV/DL, in this case is
> EPtot+RESTRAINT=VDWAALS+2*RESTRAINT, although I do NOT want to include
> the restraint energy in the dv/dl energy. In my book, if the only
> thing I'm perturbing is the van der Waals-energy, the dv/dl energy
> should be equal to the van der Waals-energy, i.e. ~8 kcal in this
> case. Am I wrong?
>
> And then about the decomposition, if I understand it correctly, the
> total VDWAALS part in the decomposition should add up to the VDWAALS
> energy in the DV/DL summary. But here I see a small difference on the
> second decimal. Is this just an numerical issue, or I'm wrong in my
> interpretation?
>
> Best regards,
> Samuel
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
Sicherer, schneller und einfacher. Die aktuellen Internet-Browser -
jetzt kostenlos herunterladen! http://portal.gmx.net/de/go/chbrowser
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Feb 19 2010 - 08:00:02 PST
