Additional materials in ISA-Tab

[Mathias Kuhring]

Dear ISA team,

I was wondering whether it is technically correct (i.e. by specification) to add own material nodes within an ISA-Tab. The ISA Abstract Model seems to be rather open for generic materials, it only defines Source and Sample as specialized material nodes. The ISA-Tab format however defines a couple of other additional specialized material nodes including Extract and Labeled Extract. The default configurations seem to feature only those specific material nodes (i.e. Source, Sample, Extract, Labeled Extract). But what if I want to model more complex procedures/experiments and like to record additional/intermediate materials with their meta data. Could I add other materials, e.g. in the form of "MaterialXY Name" (as long as it doesn't clashes with predefined materials and special cases)? Or does this disagree with the ISA-Tab specification?

Thank you in advance!

Best,

Mathias

[Mathias Kuhring]

I just realized that the mail correspondence on this topic was not feed back to the google group. So as reference, here is the rest of the discussion:

Dear Philippe,

my apologies for the late answer!

"Library Name" was in fact the specific Material we had currently in mind and it would be grate to see this one implemented.

We use ISA-Tab for integrating different omics data into a management system, including so far MS for proteomics and metabolomics as well as NGS for transcriptome, genome and exome sequencing. Organisms are currently mainly human and mice, including tissue, blood and urine samples as well as cultivated cell lines.

I was under the impression that by the original specification, the ISA model is in theory flexible enough to define own material and process nodes as needed. However, it seems the implementation in form of json and tab are restricted to a handful of specific materials. So I'm still wondering, is this an actual limitation of the json/tab specification (could find any such) or is this actually just a limitation of the API/implementation. So in theory, if I wouldn't use the API at all and just set up an ISA-Tab manually, would it be ok by specification to add arbitrary materials?

In practice of cause we would like stay compatible as much with the API as possible. So implementing a library material would be great.

Thanks a lot.

Best,
Mathias

 

 2018-06-05 18:15 GMT+02:00 Philippe:

Dear Mathias,

I wanted to check again with you regarding your cases and how to accommodate your set up.
Would you mind sending more background information about the project you are involved in (technology, sample processing needs, biological organism of interest)?

Kind regards

Philippe

 On 25/05/2018 18:24, Philippe wrote:

Dear Mathias,

Thank you for your enquiry, this is a very valid point and we have been considering it ourself.
I could imagine extension to be able to talking "Library Name' in the context of sequencing applications and so on.
While the json schema enum could be  extended, some of the ISA parsing tools would not necessarily be aware of these extensions and would there throw an error.
The issue would mainly be with the ISA-Tab conversion as we'd need to know how to deal with the new elements.
Having said that, we could discuss your use case in more details as it could possibly warrant a change or an extension of the services to behave gracefully is such extensions were made.

Let us know

Best regards

Philippe