[lammps-users] incorrect results of unsolvated protein running with LAMMPS

Dear all,

I am learning to do MD of a simple unsolvated protein with LAMMPS. My results don't make sense. Here is
how I run it.

Starting starting from the 1HE9.pdb file from pdb.org, I deleted water from it, got he9p.pdb, and then generated
an he9.pgn file which is like:
  package require psfgen
  topology top_all22_prot.inp
  pdbalias residue HIS HSE
  pdbalias residue MSE MET
  pdbalias atom MSE SE SD
  pdbalias atom ILE CD1 CD
  segment A {pdb he9p.pdb}
  coordpdb he9p.pdb A
  writepdb he9.pdb
  writepsf he9.psf

Using VMD to call the above file: vmd -e he9.pgn, I got 2 new files: he9.pdb and he9.psf.
Next I create yet another two new files needed by lammps: he9.in and he9.data by running this command:
perl charmm2lammps.pl all22_prot he9 -charmm -water=0

Now I got he9.in and he9.data. The input file he9.in is not exactly what I need, because at least I need an
implicit solvent model (pair_style is lj/charmm/coul/charmm/implicit ), so I changed it to:
units real
neigh_modify delay 2 every 1

atom_style full
bond_style harmonic
angle_style charmm
dihedral_style charmm
improper_style harmonic

pair_style lj/charmm/coul/charmm/implicit 9 10
pair_modify mix arithmetic

read_data he9.data

special_bonds charmm
fix 1 all nve
fix 2 all langevin 295 295 50 12345678
fix 3 all shake 1e-6 500 0 m 1.0 a 142

thermo 1
thermo_style multi
timestep 0.5

restart 49 he9.restart1 he9.restart2
dump 1 all atom 1 he9.dump
dump_modify 1 image yes scale no

run 100

Still the above file wouldn't give me the correct configuration, which is expected to be close to the initial native
state. I also tried to modify it with no luck. Does anybody have the knowledge about how to do it? Thank you do
much for your attention.

Xiuping Tao

Paul is planning to talk to you next week …