Re: preparing a dimeric protein for NAMD

From: Peter Freddolino (
Date: Fri Dec 23 2005 - 14:41:19 CST

Hi Irene,
the similar numbering by itself should not be a problem. However, if you
have multiple chains, you need to feed them in to psfgen as two separate
files for it to properly generate the proteins and (especially) the
terminii. If you do not do that, chances are that their connectivity has
them joined end to end. You should probably split the chains into two
pdb files, and then in psfgen do something like
segment A {pdb chainA.pdb}
segment B {pdb chainB.pdb}
coordpdb chainA.pdb A
coordpdb chainB.pdb B

Does this make sense? The chain field by itself does not make a
difference. Counterions are also very important, so you should add these
using autoionize in vmd.


Irene Newhouse wrote:

> We have a dimer protein we're working with, and we're wondering
> whether psfgen can cope with having the corresponding residues in both
> monomers having the same resid , or do we have to renumber the 2nd
> monomer?
> The pdb file from which we start [built by genomics types from an
> homology model] starts with residue 26. Could that be a problem? At
> the very right of each line is a field CHA for every atom in monomer
> 1 & CHB inr monomer 2, but the field right after the atom type, which
> should contain A or B is empty. [That matters to some viewers].
> Should I populate it with As & Bs as appropriate, or doesn't it matter
> to psfgen?
> [We thought we'd gotten through the process with the initial pdb file
> as is, but the protein sort of uncoils during MD. The other thing
> that's different from our previous projects is that this protein has
> a lot of charged residues, so needs to have counterions added. We're
> trying to figure out where things went wrong & these two differences
> are our current focus].
> Thanks!
> Irene Newhouse
> U. of Hawaii

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:41:28 CST