Define minimal metadata templates for different levels

[JolandaS]

Define minimal data, based on the transcriptomics metadata/ontologies found (see #75 ) for different levels:

• shallow (e.g. disease) - not necessarily transcriptomics specific
• data type specific (e.g. cell type)
• experiment (e.g. experiment type)

[JolandaS]

@PeterWoollard @daniwelter @karsten-quast I defined a new issue based on our discussions yesterday, to help you forward in the next step of defining transcriptomics templates. Please feel free to change/correct the description of this issue.

[mkoatwork]

Template_SummarizedDataForGRIT42_V2.xlsx

This is the template for uploading summarized AMR data into a shared Repository based on GRIT42 software. As explained on the worksheet called 'Info' some columns should be filled out by using the drop down feature. All list of values are colected in a worksheet called 'Dictionary'. Beside the list of values we have additional columns for the ontology terms. A simple example is shown in columns U and V. The gender 'Male' is resolved by the NCIT term http://purl.obolibrary.org/obo/NCIT_C16576. But in column Q it is more comlpicated as we don't have for all the bacterail strains a corresponding term in the NCBI Taxonomy. Only for the first row 'ACIBA 19606' which is Acinetobater baumannii 19606 there is corresponding URI 'http://purl.obolibrary.org/obo/NCBITaxon_575584'. For the next row we only have the species name in NCBI Taxonomy. Therefore I prosed to use a JSON structure similar to JSONLD to describe both terms 'species' and 'strain' as separate key value pairs. Thus we can use a freetext instead of a URI for all strains where no term is given in NCBI (see cell U3).
The same approach can then be used to discribe the Experiment (column A). Here we have multiple terms like 'Accumulation' and 'bacteria' which can be discribed in a similar JSONLD structure (see column B for examples). Unfortuantely here we don't have terms like 'strain' or 'species' to describe the experiment.
My questions:
Is JSONLD useful to describe combinations of terms like in the species - strain case or is there another option?
If yes, what term should we use to build the key for the key:value pair (see proposals in column B)?
How should we handle multiple ontologies like in cells B3 and B5?
Thanks
Manfred

[PeterWoollard]

Hi Manfred,
I agree, handling species and strain as two separate entities is good practice.
(and yes would like to see the mammalian organisms with scientific name + NCBI taxononomy IDs.
column B? I am seeing batch -id Noso001-1
Yes URI or URLs make sense for FAIRness

Obviously for human non-pathogen experiments, which before COVID was the majority. this template works. For other experiments, a slightly different template will be needed. It is a good exemplar though.

Thanks,
Peter

[mkoatwork]

Hi Peter, Thanks for your reply. Sorry, I was not precise: column B is on the worksheet "Dictionary" In AMR we thought of a general template for all kind of summarized data, but meanwhile we started to split into clinical and pre-clinical studies. The reason is that we don't want to have mandatory and optional columns. Optional columns will ultimately lead to incomplete filled out templates. On the other hand we would like to keep the number of templates as low as possible to not confuse users. Best regards/Viele Grüße Manfred

…

-----Original Message----- From: PeterWoollard [mailto:[email protected]] Sent: Mittwoch, 1. April 2020 10:08 To: FAIRplus/FAIRPlus_squad2 <[email protected]> Cc: Manfred Kohler <[email protected]>; Manual <[email protected]> Subject: Re: [FAIRplus/FAIRPlus_squad2] Define minimal metadata templates for different levels (#79) Hi Manfred, I agree, handling species and strain as two separate entities is good practice. (and yes would like to see the mammalian organisms with scientific name + NCBI taxononomy IDs. column B? I am seeing batch -id Noso001-1 Yes URI or URLs make sense for FAIRness Obviously for human non-pathogen experiments, which before COVID was the majority. this template works. For other experiments, a slightly different template will be needed. It is a good exemplar though. Thanks, Peter — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#79 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEXHIJCV5N3STQMOUPK33C3RKLY55ANCNFSM4LPC6DEQ> . <https://github.com/notifications/beacon/AEXHIJG2MJHP73HMVOPWKNLRKLY55A5CNFSM4LPC6DE2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEQXZT7Q.gif>