On Thu, Mar 12, 2009 at 4:41 AM, Eldon Ulrich <elu@xxxxxxxxxxxxx> wrote:
Hi,
It would just be nice to
have a fixed set of values for '.model_fit', which can be different
for each spin, to allow this to be parsable for automatic reading into
relax or other programs.
I agree completely. The '.Model_fit' tag will be enumerated with the strings
for the parameter list exactly as listed in your XML. Also included, will be
the strings for ModelFree. You will need to include a software saveframe in
your output that identifies RELAX as the software used and the version
number so software parsing the file can correctly interpret the '.Model_fit'
strings. It also may be useful to include a 'Method' saveframe where a text
config file could be included or a list of parameter names and their values
could be provided that would define how RELAX was actually run to produce
the set of output data.
Ok, I'll get to work on this. Would the Modelfree4 values for
.model_fit be the '0000110' string?
Most people will run relax in a scripted mode, and hence this script
could possibly be thrown into this method saveframe. The majority
will use
http://svn.gna.org/viewcvs/relax/1.3/sample_scripts/full_analysis.py?view=markup.
However they could have used a set of different scripts, so I could
add code to allow the user to specify these scripts. Maybe I could
put these in different method saveframes? If so, is there a way of
grouping method saveframes? Maybe use the first method saveframe to
specify how the scripts where used together, allowing the user to
input any human readable text description? I would assume these
saveframes are for humans and not computers to read.
Before I add this, I'll try to get the new order_param and
general_relaxation saveframes put together. This might take a little
while due to the saveframe reordering for general_relaxation - and
supporting this in the BMRB API I'm writing. I'll have to abstract
the relaxation data reading and writing parts even more to single
functions for all data types.
The '.Coherence_type' tag will also be enumerated with specific strings like
those that I suggested in an e-mail a few weeks ago.
This works for autorelaxation, e.g. R1, and R2 where Iz to Iz and I+
to I+ (or I- to I-) are used in the relaxation superoperator
respectively. It should be also fine for the cross correlation terms
eta_z and eta_xy where again Iz to Iz and I+ to I+ (or I- to I-) are
used in the relaxation superoperator respectively. However when you
use 2 different operators in this superoperator, for example Iz and
Sz, then you have the NOE. With degenerate transitions, more of these
cross relaxation rates (note this is not interference or cross
correlated relaxation) are observable. Despite only the heteronuclear
NOE being normally measured, would you like to design the
general_relaxation saveframe to be more general to include everything
possible (just in case someone does collect something strange). If
so, I would suggest something like this:
_General_relaxation_list.Relaxation_operator_1 Iz
_General_relaxation_list.Relaxation_operator_2 Sz
_General_relaxation_list.Relaxation_mechanisms_1 'Dipolar'
_General_relaxation_list.Relaxation_mechanisms_2 'Dipolar'
This describes the sigma_NOE or cross relaxation rate constant between
Iz and Sz. Or something like:
_General_relaxation_list.Relaxation_operator_1 I+
_General_relaxation_list.Relaxation_operator_2 I+
_General_relaxation_list.Relaxation_mechanisms_1 'Dipolar, CSA'
_General_relaxation_list.Relaxation_mechanisms_2 'Dipolar, CSA'
for the R2. This obviously doesn't include any Rex relaxation which
is in the R2. The relaxation mechanisms could include 'Quadrapolar',
'Dipolar', 'CSA', etc. Or for eta_z:
_General_relaxation_list.Relaxation_operator_1 Iz
_General_relaxation_list.Relaxation_operator_2 Iz
_General_relaxation_list.Relaxation_mechanisms_1 'Dipolar'
_General_relaxation_list.Relaxation_mechanisms_2 'CSA'
I think asking Art Palmer for an opinion about this would be a good
idea. I'm not sure if this is flexible enough and can cover all
relaxation processes, or even that it is a correct and full
description! If the cross correlated relaxation is absorbed into
general_relaxation, many of the tags from the
dipole_CSA_cross_correlations and dipole_dipole_cross_correlations
saveframe categories will probably need to be brought across to fully
specify the system. And if not absorbed, then many of these ideas of
the general_relaxation saveframe category will need to go into these
saveframe categories as well.
The 'Order_param.CSA_val' tag should be deleted. I became carried away with
inserting tags and forgot that these data would be included in a CSA
saveframe.
Ok, I will not shift CSAs. What about the bond lengths used? I
haven't found a suitable saveframe yet. Does it need to go into the
order_parameters saveframe? Also for the bond vector info (or
multiple bond vectors). And have you thought about what to do about
SRLS, RDC, etc. order parameter based analyses?
Thank you very much for including my name on your poster. I mentioned this
effort in my abstract.
Thanks. I'll make sure to make a draft of the poster available for
checking before printing.
Cheers,
Edward
Re: To all relax developers: presenting relax at the ENC conference and preliminary BMRB NMR-STAR file creation for model-free results.