Hi,
On Thu, May 31, 2012 at 3:00 PM, Dean Cuebas
<deancuebas.missouristate.edu> wrote:
> Analysis using nastruct of the first 5ns of the production run gives a
> file where some of the frames are missing all the parameter data for one
> of the base-pairs; for example all base-pair info might be present for a
> given frame except that a line for base-pair 6-19 is completely missing.
The nastruct code in cpptraj makes no assumptions about base-pairing
and looks for base pairs for each frame read in. Base-pair parameters
are only calculated for detected base-pairs. Base pairing depends on
both distance of the reference frames as well as hydrogen bonding. My
guess is that for some reason nastruct thinks that base pair is broken
for that frame.
> Since nastruct claims to give identical results to 3DNA, but I have never
> used 3DNA, I am not sure if this is normal behavior or not, and I have no
> idea on how to interpret it.
I should have been more clear here: the procedure to calculate the
actual NA base pair, base pair step, and helical parameters is the
same one used by 3DNA, but the procedure for determining base pairing
is not (AFAIK). So as long as the base pairing is the same you will
obtain the exact same numbers. I believe that for 3DNA data is only
printed for detected base pairs.
> Since sometimes NO information is given for a particular base-pair in a
> trajectory frame, and visual inspection of the frames in question in VMD
> does not show any gross abnormalities of the base-pair, I have no clue as
> to what is causing nastruct to provide no information for that particular
> base-pair.
If nastruct is not detecting that base pair during that frame it will
not print any data for it. Note that this does differ from cpptraj
previous behavior; it is an unintended side-effect of a change to the
way nastruct results are printed out. Originally I had kept everything
in memory until printout, so a missing base pair would just get 0s
written for that frames parameters. However for very large systems and
trajectories this used up quite a bit of memory, so I switched to
printing results out as they were processed, which means now missing
pairs are just skipped.
Could you send me (off-list) a topology file as well as a frame or two
where this occurs (i.e. data for base pairs is missing) so I can see
what is going on? It is possible there is a bug somewhere and if I can
reproduce it, it will be much easier to track down.
Also (or alternatively) if you compile cpptraj with '-DNASTRUCTDEBUG'
this will turn on lots and lots of debugging info for NAstruct
(specifying 'actiondebug 2' prior to the 'nastruct' command will give
some extra info as well).
Hope this is helpful,
-Dan
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu May 31 2012 - 14:00:03 PDT
