How to treat electrostatics in crystal lattice MD simulation

From: Matthew Guberman-Pfeffer (
Date: Mon Sep 07 2020 - 07:02:17 CDT

Dear NAMD community,

I am trying to run MD simulations that reproduce the crystalline
environment, as closely as possible, for a protein. I’ve defined my
simulation cell to be equal to one crystallographic unit cell, which
requires that a fraction of the protein hangs over the unit cell boundary.
If I visualize the image cells, there are no clashes between atoms. The
simulations promptly return a segmentation fault error (see below) unless I
turn off PME. I don’t understand why this is the case since a crystal
lattice should be ideal for the use of PME, unless I’m mistaken. I've
tried artificially increasing the box size so that the protein is entirely
contained within the box, but the simulation still crashed with the same
segmentation fault if PME was on. How should I treat electrostatics when
simulating a crystal lattice? If you think something else is amiss with the
simulation, please let me know as well. I’ve attached the configuration
file to this message.

Info: Finished startup at 1.25623 s, 80.1914 MB of memory in use

TCL: Minimizing for 10000 steps
Segmentation fault (core dumped)


This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:09 CST