Hi,
That's ok. Seb and I are still working out what is necessary. I
think the most important new tag would be something like:
_Order_parameter_list.Rex_frequency_1H
as then the Rex values will make sense. The Rex value changes
quadratically with field strength (this is the assumption in a
model-free analysis) so you need to specify the field strength that
the reported values apply to. Additionally Seb suggested reporting
the sigma_Rex field strength independent value in the tags:
_Order_param.Sigma_rex_val
_Order_param.Sigma_rex_val_fit_err
As:
sigma_ex = Rex / omega**2,
these last two tags are redundant. However they could be useful.
For the experimental information about the relaxation data, the
recycle delay is one of the most important parameters. Here the
NMR_spectrometer_expt saveframe could be used. Though this data
(recycle delay and interleaving order) is highly specific to each
experiment type. Is the _Spectral_acq_param.Name tag an enumerated
list? The documentation is confusing
(http://www.bmrb.wisc.edu/dictionary/3.1html_frame/frame_SpectralAcqParam.html#_Spectral_acq_param.Name).
Assuming second units, could you have something like:
loop_
_Spectral_acq_param.ID
_Spectral_acq_param.Acquisition_dimension_ID
_Spectral_acq_param.Name
_Spectral_acq_param.Val
_Spectral_acq_param.Entry_ID
_Spectral_acq_param.NMR_spec_expt_ID
1 1 'relaxation delay 1' 0.300 3452 1
2 1 'relaxation delay 2' 0.020 3452 1
3 1 'relaxation delay 3' 1.000 3452 1
4 1 'relaxation delay 4' 0.010 3452 1
5 1 'relaxation delay 5' 1.500 3452 1
6 1 'recycle delay' 5 3452 1
stop_
The names in _Spectral_acq_param.Name do not seem to be very specific
and users would probably abuse this and use very different names
depending on their background (Varian vs. Bruker, etc). Is there a
way of making sure the notation will be consistent between BMRB
entries?
Also, how would I link an Auto_relaxation saveframe to a
NMR_spectrometer_expt saveframe? Would I need to create an Experiment
saveframe (relax already does this for each relaxation data set), and
point the and Auto_relaxation saveframe to it (I don't think relax
does this correctly yet), and then the Experiment saveframe to the
NMR_spectrometer_expt saveframe via the tags:
_Auto_relaxation_experiment.Experiment_ID
_Experiment.NMR_spec_expt_ID
This seems to be becoming harder and harder to implement :S
Cheers,
Edward
On 8 March 2011 21:01, Eldon Ulrich <elu@xxxxxxxxxxxxx> wrote:
Hi,
Sorry, I was out of the office this morning and have just read through the
email string. It will take me a little time to digest this all. A couple of
general comments:
1. We have tried in NMR-STAR to avoid having tag values that take on special
formats and syntax. The _Auto_relaxation.Delay_interleaving might be handled
like this if this information was to be included with the experimental data
(see below for other comment):
loop_
_Auto_relaxation.Delay_ordinal
_Auto_relaxation.Delay_value
1 300
2 20
3 1000
4 10
5 15000
stop_
This preserves the values and from the order indicates the interleaving. A
general tag _Auto_relaxation.Delay_interleaving with values of yes/no could
be used as a flag.
2. Information like relaxation delays might be more appropriate in a save
frame where the acquisition parameters used in carrying out an experiment
were listed. The design for this in the NMR-STAR dictionary may need to be
revised or extended, but looks like this:
loop_
_Spectral_acq_param.ID
_Spectral_acq_param.Acquisition_dimension_ID
_Spectral_acq_param.Name
_Spectral_acq_param.Val
_Spectral_acq_param.Entry_ID
_Spectral_acq_param.NMR_spec_expt_ID
1 1 'relaxation delay 1' 300 3452 1
2 1 'relaxation delay 2' 20 3452 1
3 1 'relaxation delay 3' 1000 3452 1
4 1 'relaxation delay 4' 10 3452 1
5 1 'relaxation delay 5' 1500 3452 1
stop_
In this construct a tag is not needed for every parameter that needs to be
defined. The parameter name is the value to a tag and then a value provided.
As currently implemented, the order in which the parameters are executed is
not captured with an 'ordinal' tag.
If the delay information is being extracted from the acquisition files, I
would lean towards putting the information into the NMR_spec_expt save frame
and then linking the experiment to the data set save frame. If the depositor
is being asked to provide the information, then it is trickier. Something to
think about.
I think the Rex_frequency_1H information request was mentioned earlier. Was
it suggested that a tag like this should also be included in another save
frame?
Cheers,
Eldon
On 3/8/11 7:01 AM, Sébastien Morin wrote:
* If the magnetic field independent value should be stored somewhere
else. I
guess that Edward has already discussed this before in some post..?
Any idea / comment ?
I have only recently noticed this problem with the scanning of the
entire BMRB. But I think we need a tag identifying the frequency of
the Rex data in the Order_parameter saveframe. For example:
_Order_parameter_list.Rex_frequency_1H
This is modelled on the
_Auto_relaxation_list.Spectrometer_frequency_1H tag. Alternatively
the Rex values could be the sigma_Rex field strength independent
value, though none of the current data is in this notation.
I would suggest that the values are stored as both field dependent
Rex and
field independent sigma_Rex. The former would be useful for human
perception
and comparison with the amplitude of the R2 data, and should
definitely be
associated to a magnetic field (!). The latter would be better for
direct
comparison between different systems / proteins...
Then we could have the new tags:
_Order_parameter_list.Rex_frequency_1H
_Order_param.Sigma_rex_val
_Order_param.Sigma_rex_val_fit_err
And keep the old tags:
_Order_param.Rex_val
_Order_param.Rex_val_fit_err
These 5 should cover everything.
Regards,
Edward
That sounds very good.
Séb :)
Re: Rex in the BMRB