Coarse-grained simulation with Polarizable Water crashing

From: Sourav Ray (souravray90_at_gmail.com)
Date: Sun May 15 2016 - 16:34:18 CDT

Hello

I am currently trying to replace the non-polarizable water given in martini
tutorial with polarizable one in my CG simulations. I performed a 2 ns
multi-step relaxation with "langevin damping" of 10/ps and time step 10 fs,
followed by a production run with "langevin damping" decreased to 1/ps.
That simulation crashes around ~4500 step.

I keep getting error of polarizable water "atoms moving too fast" and when
I tried using "velocityQuenching", it remains fine till quenching is
applied and then shows "bad global angle/dihedral count" or "margin too
small" when switched off. Also, sequential reduction of "langevin damping"
is not working either.

None of these errors were present while using single bead non-polarizable
water or even with TIP3P water with 2 fs time step (full atomistic)
simulation with similar relaxation protocol.

The relevant additions to the martini parameter files w.r.t polarizable
water:

!!

NONBONDED
D 0.00 -1.195 2.35 ! DO NOT USE
POL 0.00 -1.195 2.35 ! DO NOT USE

NBFIX

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:25 CST