[gmx-users] Need Help on Simulated annealing (Implicit Solvation)

[Justin Lemkul]

On 1/24/13 10:26 AM, Raghuvir R. S. Pissurlenkar wrote:
> Hi
> I have tried to write a grompp.mdp file for simulated annealing of a peptide (~30 aa) in implicit solvent.
>
> Please someone could tell me if the parameters set are fine for the same.
>
> ---------------------------------------------------------------------------------
> ; Preprocessing
> define = -DFLEXIBLE
>

Irrelevant if there is no explicit water, and shouldn't be used for MD anyway 
(only EM).

> ; Run control
> integrator = md
> dt = 0.002
> nsteps = 500000; 1 ns
> comm-mode = angular
> nstcomm = 1000

grompp will complain about this.  nstcomm should be equal to nstcalcenergy.

> comm-grps = system
>
> ; Langevin dynamics
> bd-fric = 0
> ld-seed = 1993
>
> ; Output control
> nstxout = 1000
> nstvout = 1000
> nstfout = 1000
> nstlog = 1000
> nstenergy = 1000
> nstxtcout = 1000
>
> ; Neighbor searching
> cutoff-scheme = group
> nstlist = 10

Better to use a fixed neighbor list (nstlist = 0) to use the all-vs-all kernels. 
  You're using infinite cutoffs anyway, so all interactions are calculated. 
There's no point in updating a neighbor list that contains every atom in it.

> ns-type = simple
> pbc = no
> rlist = 0
>
> ; Electrostatics
> coulombtype = cut-off
> rcoulomb = 0
>
> ; VdW
> vdwtype = cut-off
> rvdw = 0
>
> ; Temperature coupling
> tcoupl = v-rescale
> tc-grps = Protein
> tau-t = 0.5
> ref-t = 300.0
>
> ; Pressure coupling
> pcoupl = no
>
> ; Simulated annealing
> annealing = single
> annealing-npoints = 1
> annealing-npoints = 0 10 20 30 40 50

You've got the same keyword twice.  The first instance (annealing-npoints = 1) 
is correct; there is only one group specified in tc-grps.

Presumably what you want here is annealing-time rather than annealing-npoints.

> annealing-temp = 300 320 340 360 380 400
>

Simulated annealing is a linear increase in temperature, and thus all you need 
to do to warm from 300K to 400K over 50 ps is:

annealing-time = 0 50
annealing-temp = 300 400

> ; Velocity generation
> gen-vel = yes
> gen-temp = 300.0
> gen-seed = 173529
>
> ; Bonds
> constraints = all-bonds
> constraint-algorithm = P-LINCS

P-LINCS is not a valid setting.  If running in parallel, LINCS automatically 
uses P-LINCS.

> lincs-order = 4
> lincs-iter = 1
>
> ; Implicit solvent
> implicit-solvent = GBSA
> gb-algorithm = OBC
> nstgbradii = 1.0
> rgbradii = 1.0

Set both nstgbradii and rgbradii to zero when using infinite cutoffs.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================