Hi,
Please see below.
On 21 February 2011 09:06, Dr. Klaus-Peter Neidig
<peter.neidig@xxxxxxxxxxxxxxxxx> wrote:
Dear Edward,
I'm back in the office, let me now reply to the various points:
Used integrals: The PDC offers 4 options. These are (and written like this
into the output):
- peak intensities
- area integral
- peak shape integral
- deconvoluted peak integral (x) where x is: lorentzian or gaussian or
fixed gauss/lorentz or variable gauss/lorentz
"peak volumes" ? Was this used in one of the older output files ?
I just assumed this name, as 'peak intensities' was in the example
files. This code was not active in relax version 1.3.10 that I
released on Friday, as it required the merger of the BMRB branch.
After releasing relax 1.3.10, I then merged the BMRB branch back into
the main relax 1.3 line. So now I have activated the code, and it is
checking for the text 'area integral' for the volume integration
method.
I was just wondering, for the last two integration methods, how are
errors estimated if no spectra are duplicated? In relax this is
forbidden as I know of no way to convert the baseplane RMSD to an
error on these intensity estimates, apart from using Monte Carlo
simulations for the peak shape fitting.
Default values: I'm not quite sure what you mean. The PDC has default
values for all settings. The the TopSpin
parameter files of the spectra are read and information is taken from there
if available. Then The user gets
everything on display and can change. Whatever he does latest, counts and
is save if the user uses the save option.
However, at the very end I do not check whether the settings correspond to
the default values or not and I do not
indicate this in the export. I think noe of the Bruker software does
something like this.
Ah, it comes out of the acqus files. Users are often lazy though, so
for example the pH value might be unchanged - the user may think this
is not used. But if we pick up this value in relax and deposit it in
the BMRB, the database will contain incorrect values. I was just
wondering if it was possible for relax to pick up these situations,
and to warn the user and drop the value from the NMR-STAR deposition
file.
Concerning the peaks used there are various options like
- automatically picked peaks
- peaks from peaklist.xml at spectrum
- peaks from xxxx where xxx is a peaklist.xml of any other spectrum
Peak positions/annotations can be done/modified in the PDC but initially
the ones contained in the loaded
peaklist.xml are used. These files also contain a tag for 'type' which can
be one out of seven types (auto,
manual, etc.). I have not imported all this information into the PDC since
I assumed the peaklists are deposit
separately into BMRB. If the user modifies peaks in the PDC he can update
peaklist.xml on disk.
The chemical shifts are deposited seperately into the BMRB. Eldon, is
there a way to associate 2 BMRB entries, so that relax can
automatically read in chemical shifts for a given model-free BMRB
deposition?
PDC info: I would guess (tentatively) it should read like:
# Protein Dynamics Center software.
SOFTWARE['rotein Dynamics Center'] = ?? // this I don't understand this
field
This line is just code to initialise the SOFTWARE dictionary (or hash)
entry to the Software_store class.
SOFTWARE['rotein Dynamics Center'].name = "Protein Dynamics Center"
SOFTWARE['rotein Dynamics Center'].authors = "Bruker BioSpin GmbH"
SOFTWARE['rotein Dynamics Center'].url = "http://www.bruker-biospin.com";
SOFTWARE['rotein Dynamics Center'].tasks = ["relaxation analysis"]
Thanks, I have added all this information and it will be included into
the BMRB deposition file. Will there be a paper/communication
presenting the PDC? I can add this in the future as needed.
Cheers,
Edward
Re: export