Re: [NAMD] How to build and simulate a periodic DNA in NAMD

From: Vermaas, Joshua (
Date: Fri Jun 09 2017 - 12:42:49 CDT

Hi Sebastian,
This sounds similar to the issues faced with periodic cellulose simulation. For cellulose, what you typically do is equilibrate at constant pressure for a little bit, but allow the motion of the periodic cell to differ along the cellulose axis and the non-cellulose axis. In NAMD parlance, this would require setting useFlexibibleCell and useConstantRatio. The initial z-axis dimension would be the length of your DNA (at least what you think it ought to be), and the x and y dimensions are set so you have enough of a water buffer. During equilibration, the DNA should sort itself out, assuming the helix is close to repeating at the length you are interested in.


On 06/09/2017 11:35 AM, The Cromicus Productions wrote:
Hi everyone, sorry to come back to this but there's something I haven't been able to solve and that is
how to work with the boundary conditions of the box. I would guess in order to simulate the periodic DNA
the first and last nucleotides must be very close and we shouldn't have any water in the middle. Also, the distances
involved should be the same than for a regular pair of base pairs. How can this be solved? Do I just consider
a slightly smaller box than my DNA and hope for the best during minimization?



On Sun, May 7, 2017 at 1:03 AM, The Cromicus Productions <<>> wrote:
I found the problem, it's not the same to add res 1 70 than 70 1 hehe

On Sun, May 7, 2017 at 12:30 AM, The Cromicus Productions <<>> wrote:
Sorry to bother again, I'm currently running into a problem of the kind "UNABLE TO FIND ANGLE PARAMETERS FOR..." the atoms that I have
just bonded. Have you ran into that problem before?

On Sat, May 6, 2017 at 8:02 PM, The Cromicus Productions <<>> wrote:
Thank you very much, Steve! I used the patch "LKNA" instead of "PBCrna" and I think it worked great!

On Sat, May 6, 2017 at 2:49 PM, Nielsen, Steven <<>> wrote:
Hi Sebastian,

Adding a "patch" to your topology file is the easiest way.
In my case, which is for the helical RNA strand in the TMV virus,
I wanted to make the RNA strand infinitely long.
My "pgn" file looks like:
package require psfgen
topology top_all27_prot_na_SON.inp
# add RNA
segment RNA {first none; last none; pdb rna.pdb}
coordpdb rna.pdb RNA
patch PBCrna RNA:1 RNA:150
regenerate angles dihedrals
writepdb x.pdb
writepsf x.psf

The key here is the patch command. My "PBCrna" patch, in the topology file, looks like:
PRES PBCrna 0.0
BOND 1P 2O3'

That's it!! The "regenerate" command generates the correct bends and dihedrals to
accompany the bond that I added. This patch adds a bond between atom "1P" on the
first RNA base and the atom "2O3'" on the last (#150) RNA base.

Note: in VMD the bond will be drawn across the box, but that is just how VMD displays
it, it really it across the PBC.


From:<> [<>] on behalf of The Cromicus Productions [<>]
Sent: Saturday, May 6, 2017 1:14 PM
To: NAMD list
Subject: namd-l: [NAMD] How to build and simulate a periodic DNA in NAMD

Hi everyone,

I'd like to simulate an infinitely long, periodic dsDNA in NAMD.
To do so I read in Nano Lett., 2015, 15 (12), pp 8322-8330 that I must
covalently bond the ends of the DNA to their periodic copies.
Any idea of which would be the easiest way to do this?

Thank you very much,


This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:21 CST