datatagsuite / schema

DATS JSON schemas
https://datatagsuite.github.io/docs/html/dats.html
Other
11 stars 8 forks source link

Proposed solution to model consent and license information #16

Closed agbeltran closed 6 years ago

agbeltran commented 6 years ago

In the consent_information_schema I followed the pattern name/abbreviation/description as per the table in the Consent Codes paper where name is an annotation, so the name (value) and if available, the IRI of the term describing it (valueIRI). In this case, the schema represents a single consent code, and materials and study group include arrays whose items are consent_info_schemas.

agbeltran commented 6 years ago

We may consider renaming consent_info_schema as consent_code_schema.

agbeltran commented 6 years ago

Descriptions fields changed.

agbeltran commented 6 years ago

Request by @altergc: change 'isDescribedInLicenses' to 'incorporatedIn'.

altergc commented 6 years ago

Here is an example of a Data Use Limitation (DUL) as it is used in dbGAP. How would you tag these properties?

"Health/Medical/Biomedical" "HMB" "Use of the data is limited to health/medical/biomedical purposes, does not include the study of population origins or ancestry."

At the moment, there is no ontology for this. They are described in a pdf: https://osp.od.nih.gov/wp-content/uploads/standard_data_use_limitations.pdf

agbeltran commented 6 years ago

This dbGAP DUL would be represented as:

(case 1)

{
 "name": {
    "value": "Health/Medical/Biomedical"
  },
 "abbreviation": "HMB"
 "description": "Use of the data is limited to health/medical/biomedical purposes, does not include the study of population origins or ancestry."
}

As not always there is an ontology term (or systematic code), I though that name was more appropriate than code.

If there was an ontology term, it would go in valueIRI, e.g.:

(case 2)

{
 "name": {
    "value": "Health/Medical/Biomedical",
     "valueIRI": "http://ontology.com/ontologyTermIRI"
  },
 "abbreviation": "HMB"
 "description": "Use of the data is limited to health/medical/biomedical purposes, does not include the study of population origins or ancestry."
}

Alternatively, we could have name as a string and use the identifier for the IRI:

(case 3)

{
  "identifier": "http://ontology.com/ontologyTermIRI"
 "name": "Health/Medical/Biomedical",
 "abbreviation": "HMB"
 "description": "Use of the data is limited to health/medical/biomedical purposes, does not include the study of population origins or ancestry."
}

but in this case, we would be adding more information than the ontology term (e.g. link to licenses) so I wouldn't use the ontology term as the identifier.

proccaserra commented 6 years ago

would the following be an option { "value": "Health/Medical/Biomedical", "valueIRI:"http://ontology.com/ontologyTermIRI" "abbreviation": "HMB" "description": "Use of the data is limited to health/medical/biomedical purposes, does not include the study of population origins or ancestry." }

which is essential case 3 with labels trying to reuse an existing pattern/motif we have from 'dats.annotation_schema', ?

agbeltran commented 6 years ago

This option would change the way we use the annotation_schema in the rest of the entities, where we say what attribute we are annotating.

agbeltran commented 6 years ago

Fixed! Thanks.

agbeltran commented 6 years ago

I believe this is ready to be merged, unless there are more comments @proccaserra @altergc ?