[gmx-users] implicit solvent concerns

[Albert Solernou]

Hi All,
I am thinking of running an implicit solvent simulation using Gromacs 
(.mdp file pasted at the end of the email). However, I am a bit 
concerned about its accuracy as I get different results for the 
Polarisation term. More specifically, the first step of an energy 
minimisation gives different values for this energy term depending on 
the number of threads/processes that I use (while the rest of the E. 
terms are exactly the same). I tried the Intel compilers + MKL, and the 
GNU compilers + ACML, as well as GMX versions 5.0.4 and 5.1-beta1. I 
could even get a segfault with the GNU compilers and ACML when using 2 
"processes" (gmx mdrun -nt 2).

The question is simple. Am I doing anything wrong? Is there any 
precaution that I should take?

Cheers,
Albert

-----------------------------------------------
; minim.mdp - used as input into grompp to generate em.tpr
integrator    = steep        ; Algorithm (steep = steepest descent 
minimization)
emtol        = 1000.0      ; Stop minimization when the maximum force < 
1000.0 kJ/mol/nm
emstep      = 0.01      ; Energy step size
nsteps        = 50000          ; Maximum number of (minimization) steps 
to perform

; Parameters describing how to find the neighbors of each atom and how 
to calculate the interactions
nstlist            = 1            ; Frequency to update the neighbor 
list and long range forces
cutoff-scheme   = Group
ns_type            = grid        ; Method to determine neighbor list 
(simple, grid)
coulombtype        = Cut-Off        ; Treatment of long range 
electrostatic interactions
rcoulomb        = 999.0        ; Calculate all long range interactions
rvdw            = 1.2        ; Short-range Van der Waals cut-off
pbc            = no         ; Periodic Boundary Conditions (yes/no)

; Implicit solvent parameters:
implicit-solvent     =  GBSA           ; use implicit solvent
gb-algorithm         =  OBC           ; use the OBC approach for the GB 
radius
gb-obc-alpha         =  1           ; parameters correspond to OBC-II
gb-obc-beta          =  0.8           ;
gb-obc-gamma         =  4.85           ;
gb-epsilon-solvent   =  78.5           ; as seen in the original paper.
sa-algorithm         =  Ace-approximation  ; include surface area 
contribution for non-polar atoms
sa-surface-tension   =  0.0054           ; default value.

-- 
---------------------------------
   Dr Albert Solernou
   EPSRC Research Fellow,
   Department of Physics and Astronomy,
   University of Leeds
   Tel: +44 (0)1133 431451