From: Irene Newhouse (
Date: Tue Jul 07 2009 - 00:25:20 CDT

I'm trying to do locally enhanced sampling of O2 diffusion through a globin.


The active part of my psfgen prep file looks like this (it's headed by a bunch

of aliases because the structure was supplied with nonstandard HEME atom

names I won't reproduce here. The sections I have not supplied DO work correctly

for normal MD):


segment A {
   first NTER
   last CT2
   pdb chaina.pdb
coordpdb chaina.pdb A
segment HA {pdb hemea.pdb}
coordpdb hemea.pdb HA
segment OA {pdb o2a.pdb}
coordpdb o2a.pdb OA
# do not bind O2 for LES
#patch PLO2 OA:1 HA:900 A:82
patch PHEM A:82 HA:900
regenerate angles dihedrals
multiply 5 OA:1:O1 OA:1:O2
writepsf aida5.psf
writepdb aida5.pdb



I get 5 copies of O2 in the pdb file. They have the same cartesian coords & they're written

a little oddly, in that all copies of the first O are followed by all copies of the 2nd, instead of

grouping the residue together. I found that on solvating & ionizing, the cartesian coords of

the copies are replaced by 0's & the B-factors altered. I solved that by replacing the incorrect

values in the solvated, ionized pdb file with the original values, after checking that the cartesian

coords of the protein had remained the same.


I pre-equilibrated the system by minimizing the water with protein, heme & O2 fixed. Then I

heated from 0 to 300K with the same set of fixed atoms. Then I did a short MD (0.2ns) with only the O2 fixed,

and finally I started the MD. It's gotten to 0.6 nsec & I thought I'd have a look at the trajectory

so far, in case something's gone horribly wrong. The first 4 copies of O2 behave nicely, but in the 5th

copy, the Os separate - as in >10A apart.


What, please, am I doing wrong?



Irene Newhouse



Insert movie times and more without leaving HotmailŪ.