OWL-based OME ontology

[nori]

Dear all,

I would like to discuss the Web Ontology Language (OWL)-based ontology introduced in the 2018 OME Annual Users Meeting (https://downloads.openmicroscopy.org/pr ... ayashi.pdf). The ontology is designed to be equivalent to OME data model. More concretely, I converted the XML-based OME schema into OWL by hand, without any data converter programs.
I have already prepared the OWL file and would like to share it in this community for further improvements.
The goal can be realisation of ontology-based data integration amongst various life-science datasets including imaging data, and machine-readable metadata that can be applied to artificial intelligence techniques.
Please let me know your opinions and suggestions.

Regards,
Norio

[jmoore]

Hi Norio,

looking forward to discussing. Do you also have a GitLab account? I've created https://gitlab.com/openmicroscopy/incubator/ome-owl for us to capture your files.

For everyone else's benefit: Norio and I discussed what it would look like, e.g., to provide a json-ld view of data in OMERO, potentially by adding the metadata directly to HTML pages.

cF https://www.w3.org/TR/json-ld/#embedding-json-ld-in-html-documents

Code: Select all

<script type="application/ld+json">
{
  "@context": "http://ome-owl/contexts/image.jsonld",
  "@id": "1",
  "name": "Example",
}
</script>


http://metadb.riken.jp/ could then scrape an OMERO instance for metadata to be indexed.

~Josh

[nori]

Dear Josh,

Thank you for preparing a GibLab project for OME-OWL discussion and development.
I have generated my account in GitLab. Please invite me to the GitLab group. I will upload my OWL file soon.

Regards,
Norio

[nori]

Dear Josh and members,

I 've just upload my owl file to the GitLab https://gitlab.com/openmicroscopy/incubator/ome-owl.
Further discussion will be held in the GitLab.

Thanks a lot,
Nori

[evenhuis]

Hi Josh,

we are also interested in having a JSON-LD view of the OME-schema. We would like to map metadata keys to URLs that define the terms.

What we are trying to do is collect the metadata associated with image files in JSON-LD format which then gets rendered as a webpage:

Code: Select all

{
    "@context": {
        "@vocab":"https://schema.ome.org/",
        "ome": "https://schema.ome.org/",
        "LightSource": "schema:LightSourceGroup"
    },
...       
   {
            "@id": "control/control2.nd2",
           ""omeroId": 6249,
            ....
            "@type": "File",
           "LightSource": "Lumencore"
   }


A goal is to link the keys their OME definition. For example, in the above the key "LIghtSource" would link to "https://schema.ome.org/LightSourceGroup".

The OME-schema documentation is fantastic,
https://www.openmicroscopy.org/Schemas/ ... 6/ome.html
however due to the frame is it not possible to link to a single entry.

Is it possible to autogenerate individual pages for the elements in schema?

Cheers,

Chris

[jmoore]

Hi Josh,

Hi Chris,

...

A goal is to link the keys their OME definition. For example, in the above the key "LIghtSource" would link to "https://schema.ome.org/LightSourceGroup".

The OME-schema documentation is fantastic,
https://www.openmicroscopy.org/Schemas/ ... 6/ome.html
however due to the frame is it not possible to link to a single entry.

Is it possible to autogenerate individual pages for the elements in schema?

With that particular generator (oxygen) it's not, but certainly having individual pages for each element would be doable. Nori may have further ideas, but I wonder if if we could make use of the naming schema in his issue on gitlab. E.g. http://www.openmicroscopy.org/rdf/ome_c ... ightSource would be more than just a name, it would actually resolve.

To some extent, this forum post as well at the https://gitlab.com/openmicroscopy/incubator/ome-owl repository is for gathering the community's interests and requirements. Could you expand on what you'd like to see and let's see if we can make some steps to bring all this together?

Thanks all,
~Josh

[mtbc]

Just to add, note that one can link just fine to https://www.openmicroscopy.org/Schemas/ ... ourceGroup etc.

Cheers,
Mark

[wmoore]

we are also interested in having a JSON-LD view of the OME-schema.

Isn't that exactly what is provided by the OMERO JSON API? https://docs.openmicroscopy.org/omero/5 ... n-api.html

E.g. JSON-LD data (based on OME model) for a single Image: https://docs.openmicroscopy.org/omero/5 ... ngle-image

Will.

[jmoore]

Will,

except http://www.openmicroscopy.org/Schemas/OME/2016-06#Channel doesn't resolve to anything unique, but agreed that the JSON-API should be brought included in the requirements and brought in line with any other efforts.

Cheers,
~Josh

[evenhuis]

Hi Will,

Thanks for the quick reply.

Defining the terms

Isn't that exactly what is provided by the OMERO JSON API? https://docs.openmicroscopy.org/omero/5 ... n-api.html
E.g. JSON-LD data (based on OME model) for a single Image: https://docs.openmicroscopy.org/omero/5 ... ngle-image

That is a JSON-LD description of the image. A feature of JSON-LD is "@context" which allow the terms (the keys or right hand side) in the JSON-LD to be defined by resolving to a URL.

As pointed outed by Josh, linking to
http://www.openmicroscopy.org/Schemas/O ... 06#Channel
doesn't actually take you that section and it's a huge file (~1M lines of html). For the @context to work it term in the schema needs separate page.

While not essential, it is a nice to be able to package together a dataset with the model definitions and the @context tool allows the datasets from different repos, which would different terms/keys, to mapped onto a common set of terms.

Splitting the documentation
Oxygen appears to have the option to split the docs into separate files. Search for "Split output into multiple files" on this page:
https://www.oxygenxml.com/doc/versions/ ... iple-files

If this feature isn't available due to pricing tiers of something we may be able to hack together a python script to do this.

Cheers,

Chris