From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Sun Dec 13 2009 - 15:08:59 CST
I wonder whether for constant-volume coarse-grained minimization the
force field and force-field parameters are sufficiently set with:
# Input
paraTypeCharmm on
parameters /working_dir/rbcg-2007.par
# Force-Field Parameters
exclude 1-2
cutoff 35.0
switching on
switchdist 20.0
pairlistdist 100.0
hgroupcutoff 0
I am looking for the causes that:
# Minimization
if {1} {
minimize 1000
reinitvels $temperature
}
numsteps 10; # 300 ps 0.3 ns
systematically stops at step 536 for (I guess) too high potential energy:
WRITING COORDINATES TO DCD FILE AT STEP 535
PRESSURE: 536 4.37238e+07 -3.65202e+06 -603131 -1.20215e+06
5.36159e+07 1.62008e+06 -1.70818e+06 9.75259e+06 4.9599e+07
GPRESSURE: 536 4.37238e+07 -3.65202e+06 -603131 -1.20215e+06
5.36159e+07 1.62008e+06 -1.70818e+06 9.75259e+06 4.9599e+07
ENERGY: 536 1474.2549 6309.1780 11699.6826
0.0000 -2384.6469 187117483.7999 0.0000
0.0000 0.0000 187134582.2686 0.0000
187134582.2686 187134582.2686 0.0000 48979559.0521
48979559.0521 1178217.0000 48979559.0521 48979559.0521
Although superimposition of beads comes first to mind, the cg system
(protein+bilayer) is free of clashes, and was solvated with cg water.
I come from experience with Amber full atoms, where clashes are
eliminated by steepest descent + conjugate gradient minimization. Is
the minimizer of NAMD working differently or cg solvation has
introduced too hard a situation?
If I interpret correctly, the energy was not so high in the first
steps of minimization:
MINIMIZER SLOWLY MOVING ATOMS WITH BAD CONTACTS DOWNHILL
PRESSURE: 2 3.45454e+12 1.69039e+12 -6.73379e+10 1.38241e+12
8.29396e+11 -4.76509e+10 -2.18147e+11 -2.94639e+11 1.62733e+11
GPRESSURE: 2 3.45454e+12 1.69039e+12 -6.73379e+10 1.38241e+12
8.29396e+11 -4.76509e+10 -2.18147e+11 -2.94639e+11 1.62733e+11
ENERGY: 2 1413.1771 6292.4622 11632.7598
0.0000 -2392.2611 9999999999.9999 0.0000
0.0000 0.0000 9999999999.9999 0.0000
9999999999.9999 9999999999.9999 0.0000 9999999999.9999
9999999999.9999 1178217.0000 9999999999.9999 9999999999.9999
thanks
francesco pietra
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:36 CST