Re: [Psidev-pi-dev] PSI XML <-> Ontology Mapping File

[David C. <dc...@ma...>]

Thanks, Luisa that's quite clear.
(I've now sent this message to the list for general input/thoughts)

 From a previous message, Luisa said:
 > In general I think number of those terms should be XML attributes in
 > the schema (like 'sample id', or 'date / time search performed' or
 > 'modification position', whereas CV are fine and should be the
 > reference for descriptif information like 'database filtering' or
 > 'search engines scores'.

For the record, one thing that we agreed about in Lyon was that if 2 or 
more search engines supported a particular parameter, then we'd like 
this as a 'node' in the schema.

So, for example with an MS-MS tolerance should this be CV? This 
tolerance is something that all search engines require (or maybe one day 
they estimate it, but we still want to know what value is used). So it's 
something that is 'required' and is not 'descriptive'.
But a tolerance is no use without units, so this can't be a simple 
attribute.
Similarly, in Luisa's example above: date / time search performed 
possibly needs a time zone, which would also be CV?

How should we model this? Suggestions from anyone please - or just tell 
me what's already been decided elsewhere and we'll follow that.

Another example: mass values can be calculated as 'monoisotopic' or 
'average'. This again is required for all search engines and again seems 
like it could be cv to me. Or should we use and xsd:enumeration for 
something like this where there are only 2 possibilities?


David


Luisa Montecchi wrote:
> Hi Phil,
> 
> through the mapping you can limit the unit one can use in a given 
> *schema location* (Xpath), whereas you will need to create a so called 
> 'object rule' in the validator to verify that for each *CVparam term*, 
> only an appropriate subset of units are associated with it.
> 
> In other words, the mapping allows the discrimination of the various 
> CVparam element in the schema by their Xpath and permits to restrict the 
> subset CVparam terms and/or unit terms that can be used in each location.
> 
> Dependencies between CVparam terms and unit terms cannot be encoded in 
> the mapping, but can be checked via the validator tool,
> 
> I hope this is clear,
> 
> Best regards,
> 
> 
> Luisa
> 
> 
> 
> 
> Phil Jones @ EBI wrote:
>> Hi Luisa,
>>
>> Can you confirm for me please - does the mapping file include the 
>> ability to
>> mandate the presence of particular units for specific CV term usage in an
>> XML file? (I am thinking now about mzML files, that include the unit
>> ontology accession and term in the CvParam entry - we wish to use the 
>> same
>> XML structure in analysisXML).
>>
>> Best regards,
>>
>> Phil.
>>

-- 
David Creasy
Matrix Science
64 Baker Street
London W1U 7GB, UK
Tel: +44 (0)20 7486 1050
Fax: +44 (0)20 7224 1344

dc...@ma...
http://www.matrixscience.com

Matrix Science Ltd. is registered in England and Wales
Company number 3533898