Re: Alchemical transformations and DDG of point mutations

From: Gianluca Interlandi (
Date: Fri Dec 14 2012 - 19:07:00 CST

I found that I DO need to use alchemify explicitly if I use psfgen instead
of the mutator plugin. I am using NAMD2.9 and during minimization it
complained of not funding angle parameters for CBA-CA-CBB. This is because
the PSF generated by psfgen included that angle. So, I ran alchemify on it
and minimization is running fine now.

The alchemify website says (

   "As of NAMD 2.7b2, alchemify is not needed anymore, since all unwanted
    terms are discarded by NAMD at runtime."

This does not seem to be true, at least not for minimization. Or, could it
be that I'm using a binary compiled with the memory optimized option? How
do I find that out? I'm using "NAMD_2.9_Linux-x86_64".

In any case, I ran alchemify also on the solvated and ionized structure
just to be sure that the nonbonded exclusion list is created and taken
into account.



On Thu, 13 Dec 2012, JC Gumbart wrote:

> This is irrelevant, because Mutator plus Alchemify make sure that 
> the correct exclusion list is established so that the end states of your
> alchemical transformation never see each other.
> This is outdated - recent versions of NAMD (for at least the last couple years?)
> automatically generates the exclusions without the need to add them to the psf.  I
> don't believe there's any need for Alchemify anymore.
> On Dec 13, 2012, at 12:47 PM, Chris Chipot wrote:
> Gianluca,
> 1) Is it better to use alchemy in NAMD in the FEP or TI mode
> for this type of calculatons?
> Contrary to several extraordinary claims (which have never been
> supported by extraordinary evidences), there is no real advantage
> to use TI over FEP, or the other way around. I would, nonetheless,
> advocate FEP for practical reasons -- you can subsequently use
> parseFEP for your post-analysis, notably get the BAR estimator.
> 2) You describe that in each FEP calculation, 30 intermediate
> states were considered. Does this mean that the progress of
> lambda was broken down into 30 windows each of 150+150 ps?
> How many FEP simulations did you run in total?
> Yes to the first question. I do not quite get the second -- the windows
> are chained and the net free-energy change is the sum of free-energy
> changes in individual strata.
> 3) Do I need to use AlchDecouple on or off?
> I strongly advocate AlchDecouple off, in particular for solvation
> free-energy calculations, as one simply cannot assume that the
> conformational ensembles in vacuum and in condensed phase
> are identical.
> 4) The VMD mutator plugin replaces both the wild-type and
> mutant side chains such that both clash with the backbone. If
> this is a problem, I might create the PDB myself and still
> use the PSF generated with VMD.
> This is irrelevant, because Mutator plus Alchemify make sure that
> the correct exclusion list is established so that the end states of your
> alchemical transformation never see each other.
> Sorry, lots of questions. I would appreciate any help on
> this!
> Hopefully, I was convincing enough that it will stop here!
> 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:
> Web:
> The light shines in the darkness, and the darkness has not overcome it.
> John 1:5.
> _______________________________________________________________________

Gianluca Interlandi, PhD
                     +1 (206) 685 4435

Research Scientist at the Department of Bioengineering
at the University of Washington, Seattle WA U.S.A.

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:22:22 CST