Hi Robert,
A few obvious issues that eginput/dna_refi/refine_full.py would have
with a DNA-RNA hybrid:
1. Unless you were careful with your selections, the NCS symmetry
term will fail during setup
because the two equivalent groups (the DNA strand and the RNA strand)
have different numbers
of atoms.
The solution for this would be to select everything except the 2' H
or OH groups.
2. Another problem would be in the setup of the residue-residue
position database term. Since the
code in eginput calls the DNA version, the setup script will fail
(perhaps without giving any visible
warning) when it encounters a URA residue type.
There's no perfect solution for this issue. There are separate
potential surfaces for DNA and RNA,
and since the number of DNA-RNA hybrids in the PDB is rather small,
there's no practical way to create
one for the hybrid.
That said, if you think that the DNA-DNA residue position database
would be reasonably accurate (and
it probably would be), you could trick the setup script into running
by temporarily changing the residue
type of your Us to Ts:
xplor.command(""vector do (resn = "ADE") (resid 14 or resid
21)"") # if resids 14 and 21 are Us
xplor.command("@dna_positional.setup")
xplor.command(""vector do (resn = "URA") (resid 14 or resid 21)"")
Note that the torsion angle database term is insensitive to DNA vs
RNA issues, so it should be OK.
Hope this helps.
--JK
On Jul 2, 2007, at 2:57 PM, <brinson at umbi.umd.edu> wrote:
> Dear users of Xplor-nih,
> I am new to python and xplor and am trying to compute
> a 20bp RNA-DNA hybrid. I have included the standard restraints -
> base/backbone NOEs, hydrogen-bond NOEs, dihedrals, and one set of
> RDCs. I created a decent starting structure using the anneal.py
> protocol. When I input that structure into refine_full.py from /
> eginput/dna_refi, the helix seems to become overwound in the middle
> and underwound at the ends. I have tinkered with the force
> constants and temperature (from 3000 down to 300). I added some
> "anti"-noes (~20A at various locations, generated from InsightII)
> to try to help, but my output structures did not improve. There are
> no apparent errors in the output file. I would appreciate advice
> on refining the structure.
>
> Thank you,
> Robert
> _______________________________________________
> Xplor-nih mailing list
> Xplor-nih at nmr.cit.nih.gov
> http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
