Re: [cml/ccml-discuss] conditions/metadata in CMLSpect

[peter murray-r. <pm...@ca...>]

At 10:32 01/06/2006, Stefan Kuhn wrote:
Thanks very much - sorry for the delay...

There are some important general points=20
here.  CML does not directly provide markup for=20
most chemical concepts, but enables users to=20
create dictionaries describing to concept. These=20
dictionaries can have any of the dataTypes that=20
CML allows - scalar, matrix, array, substance,=20
condition, parameter, etc. For example Bruker=20
often include several hundred concepts in their=20
file and I would expect a Bruker-NMR dictionary=20
to manage them. Most of the concepts below can be managed in this way.

The exceptions are:
* when a concept has a new data type. This might=20
happen for (say) 3- dimensional data, 3rd rank tensor, etc.
* when the data are fundamental to the=20
algorithmic processing and interpretation of the=20
information. I am not sure that there is anything=20
required for spectra. For example NMR frequency -=20
which is required to convert between Hz and ppm -=20
is only required for NMR spectra. We need to=20
avoid vocabulary pollution. Here there is a real=20
need for an NMR dictionary which contains core concepts.

>Hi all, hi Peter, we (Tobias and me) come across=20
>some problems with metadata (or sort of) in=20
>CMLSpect, starting from what is defined as=20
>conditions in NMRShiftDB. I try to translate=20
>them to CML entries. The conditons are: -=20
>Temperature: Fine, can become an entry in conditionList

Yes.

>  - Field strength: Similar, only problem:=20
> cmlDict does not seem to have an entry for=20
> field strength, similar to temperature.

Correct. Please add it! The idea of the=20
dictionaries is that they can be created by the=20
community and this looks like an excellent example.

>Or did I miss this. - Solvent: Is this a condition?

I suggest substanceList/substance. This is what=20
we do in reaction and several of your concerns are similar

>  It does not really have a unit. In CMLReact,

>  substance is used for solvents. solventList is=20
> currently not an entry in the schema for CMLSpect. Should we have it?

we should certainly have substanceList - this=20
could contain things like TMS as well

>  Are there other substances for spectra than solvent?

That's up to you. anything that is important

>Or should it go in metadata? - Assignment method=20
>(COSY, HMBC etc): This would be a metadata, right?

Probably.

>  Or is metadata more for non-chemical stuff=20
> like "measured by", "which lab" etc.?

Have a look at the comp chem examples. I think there are a lot of=
 similarities

>  Ok, I hope this feedback has some value.=20
> Stefan -- Stefan Kuhn M. A. Cologne University=20
> BioInformatics Center=20
> (http://www.cubic.uni-koeln.de) Z=C3=BClpicher Str.=20
> 47, 50674 Cologne Tel:=20
> +49(0)221-470-7428   Fax: +49 (0) 221-470-7786=20
> My public PGP key is available at http://pgp.mit.edu



>=20
>_______________________________________________=20
>cml-discuss mailing list=20
>cml...@li...=20
>https://lists.sourceforge.net/lists/listinfo/cml-discuss

Peter Murray-Rust
Unilever Centre for Molecular Sciences Informatics
University of Cambridge,
Lensfield Road,  Cambridge CB2 1EW, UK
+44-1223-763069=20