Re: Inconsistencies in the v2.1 (or v2.1.1) BMRB model-free entries. -- February 03, 2011 - 13:25

Hi Eldon,

I've now implemented support for reading the molecular_polymer
saveframe of NMR-STAR v2.1.  I have assumed that this is equivalent to
the entity saveframe in v3.x.  This is pulling out some more sequence
problems in the v2.1 files.  I'll send these inconsistencies as a diff
once I have them all sorted out.

Regards,

Edward



On 2 February 2011 14:35, Edward d'Auvergne <edward@xxxxxxxxxxxxx> wrote:
> Hi Eldon,
>
> I don't know if this is the best channel for this information.  Is
> there a BMRB mailing list where it would be better for this
> information?
>
> Ok, this is how I have found these inconsistencies.  I have used relax
> to read in the BMRB NMR-STAR formatted files.  This uses bmrblib which
> I wrote (http://gna.org/projects/bmrblib/).  This library is pretty
> close to complete for relaxation data and model-free data, and would
> be very easy to extend to handle the entirety of the NMR-STAR
> dictionary.  It can both read and write valid NMR-STAR formatted files
> in versions 2.1, 3.0, and 3.1 (a little debugging might be still
> required, and expansion to different revisions such as 2.1.1 is also
> possible).  This Python library is an abstraction of the underlying
> file format.  The very low level reading and writing of the STAR
> format is handled by Jurgen F. Doreleijers' pystarlib (jurgenfd att
> gmail dott com, http://code.google.com/p/pystarlib/).
>
> For reading the entire BMRB model-free data content, I have performed
> the following.  I have downloaded all of the files from
> http://www.bmrb.wisc.edu/search/query_grid/query_1_46.html using the
> link http://www.bmrb.wisc.edu/ftp/pub/bmrb/compress/query_1_46.tar.gz.
>  These are all in the version 2.1 or 2.1.1 format.  Then using the
> file names, I have downloaded all of the corresponding v3.1 files from
> http://www.bmrb.wisc.edu/ftp/pub/bmrb/entry_lists/nmr-star3.1/.  It
> looks like maybe 30% of the old formatted files have been converted to
> the newer format so far.  I will write 2 subsequent emails with
> explanations of the problems with the version 2.1 files and the 3.1
> files separately.
>
> In this mail, I would like to describe general problems.  The first is
> that pystarlib cannot handle the semi-colon notation in non-free
> looping tag categories, e.g.:
>
>  loop_
>     _Vendor.Name
>     _Vendor.Address
>     _Vendor.Electronic_address
>     _Vendor.Entry_ID
>
>     _Vendor.Software_ID
>
>     'J. Patrick Loria' .
> ;
> http://xbeams.chem.yale.edu/~loria/
> patrick.loria@xxxxxxxx
> ; 15097 1
>
> This is in the v3.1 file bmr15097.str.  The basic pystarlib
> functionality probably needs to be fixed, assuming this construct is a
> valid STAR format.  The second is that the bmr4970.str entry is not
> parsable.  This file has multiple 15N S2_parameters saveframes:
>
> save_S2_parameters_15N_22C
> save_S2_parameters_15N_35C
> save_S2_parameters_15N_47C
> save_S2_parameters_15N_60C
> save_S2_parameters_15N_73C
>
>
> But these all have:
>
>   loop_
>      _Sample_label
>
>      $sample_one
>
>   stop_
>
>   _Sample_conditions_label    $sample_conditions_one
>
> They might be the same sample, but the sample conditions are different
> as the temperature is changing.  By eye, this is obvious, but for the
> automatic parsing of this data, the file has to be blacklisted and
> skipped.
>
> Cheers,
>
> Edward