From: B.W.J. Irwin (bwji2_at_cam.ac.uk)
Date: Tue Jan 10 2017 - 09:09:52 CST
This .pdf is useful (sorry if you already know this)
Particularly the heading:
Why a soft-core potential should always be included
If you haven't included one.
The fepout file should show the contribution from each λ-window "Delta
Delta A" and a measure of uncertainty.
If a particular window has a large contribution you could try splitting
it into two smaller windows.
On 2017-01-10 12:15, Hannes Loeffler wrote:
> It would be useful it you kept the discussion in one thread.
> Your attachment suggests that you plot λ vs <U> but your really need
> <∂U/∂λ>. Is that what it is? How have you extracted this?
> Maybe you send all relevant input files (control file, parmfile,
> coordinates/alchfile) and output files (alchOutFile) so that somebody
> the list can take a look.
> On Tue, 10 Jan 2017 16:39:01 +0530
> Abhishek Kumar Singh <bo13m1002_at_iith.ac.in> wrote:
>> I have alchemically transformed K+ to Na+ at the center of water
>> boxes (40, 60 and 80 A edge length respectively) using NAMD software
>> and CHARMM36 forcefield. A harmonic spring of force constant = 10
>> kcal/mol.A^-2 has been applied throughout the transformation. Temp
>> and Pressure has been kept at 310K and 1 atm respectively.
>> Free energy of alchemical transformation has been calculated using
>> thermodynamic integaration. The derivative has been computed as
>> (∂G/∂λ) = ⟨ U(λ=1)- U)(λ=0) ⟩λ, where λ is the coupling parameter
>> such that λ=1 and λ=0 are K+ and Na+ vdw parameters respectively.
>> Free energy derivatives has been computed for 11 λvdw values ( 1.0,
>> 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1 and 0.0). The area under
>> the plot (derivate vs λ) gives me = -23.40, -23.46 and -23.4
>> kcal/mol for the box sizes of 40A, 60A and 80A respectively. I am
>> attaching alongwith the derivate vsλ values for 40A water box.
>> The standard CHARMM36 prm file actually states that the value should
>> be -18.6 kcal/mol.
>> So what can be reason for this difference. Is there any problem in
>> this procedure ?
>> Your suggestions are most welcome.
This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:58 CST