[gmx-users] Fwd: Relative free energy perturbation

[Davide Bonanni]

Hi,

I am trying to perform an alchemical free energy perturbation with gromacs,
since I am approaching this technique for the first time. So, it would be
so nice if you could help me to clarify some doubt.

I have built my system, which is composed of 1 protein, 2 cofactors, and 1
ligand that should changes from state A to B.
I used FESetup to make the topology for my ligand. The change I want to
apply to my molecule is just the transformation of a hydrogen into
chlorine, so I do not have dummy atoms involved. This is the only string
that changes from state A to B in my ligand topology:

[Atoms]
; Nr type resnm atom cgnr   charge    mass  typeB  chargeB   massB
   30 h4 1 LIG     H30  30   0.158051 1.0080   cl     -0.078297   35.4500

So, I have the complex.gro and complex.top, I made some minimization tests
and everything seems good. I have some doubts on the procedure I'm supposed
to follow to perform FEP calculation. Some questions:

1) Can I perform the calculation in a single step with soft core potential
enabled? I mean, is it correct to transform directly the hydrogen into a
chlorine instead of using 2 topologys and 2 complexes, where in the first
step I transform the hydrogen into dummy atom, and in the second I
transform the dummy atom into chlorine. Using soft core potential I can
perform any kind of transformation in single step, or are there some cases
where I need to use different procedures?

2) Referring to BevanLab Tutorial 6: Free Energy Calculation (
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/gm
x-tutorials/free_energy/index.html), I have to perform every step of
molecular dynamics at every Lambda value, is that right?

3) I run a test minimization step (mdp file attached) of my complex at the
last "init-lambda-state", 15 in my case. Looking at the .trr output file I
can see that the bond between the carbon and the hydrogen which should be
trasformed is longer than a normal C-H bond, but the atom is still
recognized as hydrogen (picture "http://tinypic.com/r/2wp11dh/9": purple ->
init-lambda-state = 0 ; blue -> init-lambda-state = 15). I was wondering if
this is what I am supposed to see, so if gromacs is considering the state B
of my system where I have chlorine bound to carbon instead hydrogen.

I am apologize if my questions are technically incorrect. As I said it is
the first time I'm approaching to free energy perturbation and every kind
of help and suggestion is really appreciated.

Thank you very much in advance.

Best regards,

Davide Bonanni

----------------------------------------------------------
Davide Bonanni, PhD student
Dept. Science and Drug Technology
University of Turin (UniTO) - Italy
Mobile +39 3409726272 <+39%20340%20972%206272>
----------------------------------------------------------