Re: [Psidev-ms-dev] mzML 0.99.0 submitted to document process

[Lennart M. <len...@gm...>]

Hi Andy,


> The decision about how to implement CV terms is pretty important and we should try to come up with a coherent policy across PSI if possible. Here are my thoughts:
> 
> A while back Luisa and myself drafted a proposal for mapping model elements to CV terms that may simplify some of the problems currently being worked through. The draft and sample instance are here: http://www.psidev.info/index.php?q=node/159 (see Mapping between exchange schema and CVs).
> 
> I would strongly vote for option A, and in addition maintain a mapping file. This is more work for the CV coordinators (but hopefully can be mainly automated), and would force software implementers to interact with the CV WG when they need new terms, but given the heavy reliance on CV terms in the mzML schema I see no way around this. 
> 
> If a mapping file is kept updated in parallel to the CV, software can check whether a valid term has been provided for a particular model element. In the example of spectrumType, the mapping file would specify that only child terms of spectrumType are allowed (e.g. for the model element fileContent). If a vendor publishes a file with:
> 
> <fileContent>
> 	<cvParam cvLabel="MS" accession="MS:9999999" name="SRM spectrum" value=""/>
> </fileContent>
> 
> This would automatically be rejected by the validator (or at least a warning output), as it should be, since there's no point having a CV where the terms are not controlled!  

That mapping file is effectively in use by our mzML semantic validator, 
for exactly the reasons you outlined above!
So yes - this has been made available in the larger mzML kit and has 
also been implemented online (your above example indeed does not validate).


Cheers,

lnnrt.