Hi all
I finished running the 50 ns MD Simulation of a very large system (17 Million atoms).
It is a capsid in water. I used charmm FF for the protein and spc for the water.
I am studying temperature triggered structural transitions.
Since Charmm FF is build based on TIP3p water model, are my simulation results wrong?
Is there any main issue that I should know?
Thanks
John
Hi all
I finished running the 50 ns MD Simulation of a very large system (17
Million atoms).
It is a capsid in water. I used charmm FF for the protein and spc for the
water.
I am studying temperature triggered structural transitions.
Since Charmm FF is build based on TIP3p water model, are my simulation
results wrong?
there is no absolute "right" or "wrong" in classical MD simulations;
there is only "more wrong" or "less wrong".
in fact, even the regular TIP3P model would be inconsistent, CHARMM
uses a variant of TIP3P (sometimes referred to as TIPS2) that has LJ
terms on the hydrogens), which the regular TIP3P does not have (and
neither has SPC). the LJ terms are needed to rebalance the solvation
strength.
using the SPC model with CHARMM parameters is a bad idea. i have no
data for *how* bad it is. i am certain that people have tried running
this combination and there may even be some that have been able to
publish the resulting data. but is suspect, that you won't be able to
get your work published in a respectable journal without extensive
validation, which in turn will be difficult given the subject of your
work. i would expect people to be very concerned about inconsistent
parameters for a topic that is so sensitive to subtle differences.
with that in mind, there is another issue to consider: the outdated
CHARMM implementation in LAMMPS, which does not support CMAP
corrections for proteins. so you have a double threat there. so even
if you had used the correct water model, people might be very
concerned about the validity of your simulations.
sorry, but this project sounds a lot like a job for NAMD (or Gromacs),
axel.