On Sun, Jan 26, 2014, Robyn Ayscue wrote:
>
> I'm attempting to build a additional distance geometry bounds databases for
> NAB with the hope that I can enable NAB to construct 3D oligos with
> non-canonical substructural motifs.
You are traversing an area (buidling nucleic acids using distance geometry)
that has not been well-explored in the past. I apologize if the manual seems
to indicate that this is a solved problem. It might help (but maybe not) to
know what sorts of non-canonical structures you are trying to generate.
> Using the pre-made "bdna", "arna", etc
> stacking and pairing databases in $AMBERHOME/dat/dgdb for our non-standard
> oligo foldings seems to result in 3D structures that tend to want to "kink"
> sharply, curving the overall structure into a sort of "C" shape that
> subsequent energy refinement does not rectify.
A guess: the base-pair restraint databases don't actually have enough
data about the backbones to prevent bending or kinking there. You may
have to play with the "useboundsfrom" command with backbone fragments
to get reasonable starting configurations. If you can identify some
short interatomic distances that allow the kink to form, you might be
able to add some addtional distance bounds that would straighten out the
structures.
> I get no errors printed to stdout when I run the
> "make_databases" script, but I end up with only empty files.
Again: apologies. I'm not sure any one has ever used the "make_databases"
script since Neill first wrote it 15 years ago. In my (limited) experience,
I have mainly used the "useboundsfrom()" routine, and have never tried to
create new bounds databases.
>
> Being a NAB novice (and not much of a programmer in general), it may be
> that I am completely barking up the wrong tree by attempting to build new
> bounds databases, and that instead there is something lacking in my .nab
> input files that would resolve the "kink" in my NAB-generated structures
> without going to such extremes. One of my .nab files is attached below
> (called "test_2l5k_pred.nab"), and should generate a 3D reproduction of the
> first frame of pdb code 2L5K (from pdb.org).
I can't test your script directly, since the "dna_nab10.lib" and the
"apt.nucleotide.db" files seem(?) to be ones that you have created -- I don't
see them in the Amber distribution.
In general, distance geometry is (as you have found!) a tricky way to build
nucleic acid structures. This wasn't so clear when we started the NAB
project. You might want to look at the DNA construction utilities at
http://w3dna.rutgers.edu.
....good luck....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jan 28 2014 - 05:30:03 PST
