AMBER Archive (2002)

Subject: FEP - ligand to nothing (or water)??

From: Aldo Jongejan (jongejan_at_chem.vu.nl)
Date: Tue Oct 29 2002 - 03:50:54 CST

Dear all,

Maybe what I am trying is not possible in AMBER, but I
would like to do a bit of 'computational alchemy', i.e
having a ligand transform in water molecules within
the pocket of the protein. And calculate the free energy
change.
I tried to do this in the GIBBS module, but could not
get the topology changes correct. So, I decided to
have the ligand 'dissapear' (much like the methanol in
the water box example on the Amber tutorials). I have
the dummy atoms assigned, but after only 0.08 change
in lambda, the run crashes with the following error:

 Current Lambda = nan
 Last F.E. update: Lambda = 0.080000 Step = 83992 Method =
F.E.P.
 Accumulated "forward" quantities (Regular)
    Lam+d_lam =nan F_energy = inf
    ELEC = nan NONB = nan 14NB = nan
    14EL = nan BADH = nan
    EPOL = 0.000 E3bod= 0.000
 Accumulated "reverse" quantities (Regular)
    Lam-d_lam = 0.075000 F_energy = inf
    ELEC = nan NONB = nan 14NB = nan
    14EL = nan BADH = nan
    EPOL = 0.000 E3bod= 0.000

------------------------------------------------------------------------------

       DETERMINANT IS ZERO IN CENMAS ... STOP

     INPUT INCONSISTENT IN SUBR. CENMAS

I suspect this has something to do with the fact I let
the force constants go to zero? I don't perform any
bond-shrinkage.

Does anybody have experience with this, or an alternate
solution for my problem? Or is what I am trying to
achieve impossible? Are there any other methods to
reach the same goal?

many thanks,

aldo 

-- 
###########################################

Aldo Jongejan
Molecular Modeling Group
Dept. of Pharmacochemistry
Free University of Amsterdam
De Boelelaan 1083
1081 HV Amsterdam
The Netherlands

e-mail: jongejan_at_chem.vu.nl
tlf:    +31 (0)20 4447612
fax:    +31 (0)20 4447610

###########################################