Map EGA metadata to semantic models

wna-se commented 4 months ago

The (meta)data from the Rare Disease Synthetic Dataset (EGAD00001008392) is not encoded in a format that is readily accessible for queries through the virtual platform. To the degree that it is possible and relevant to the examples of analyses being worked on, the (meta)data should be translated to corresponding elements, values and structures in the EJP-RD data models:

But also the Fairgenomes-semantic-model should be considered as a complement.

Working documents to describe mappings from EGA’s metadata schema:

[x] analyses
[x] dataset
[x] experiments
[x] runs
[x] samples
[x] studies

Cross-linking between different types of entities (records):

Each of the mappings above should be used to create executable RML-mappings that can translate the data, as illustrated in the example to FAIR genomes genomed below (using YARRRML syntax):

prefixes:
  ega.dataset: https://www.ebi.ac.uk/ega/datasets/
  NCIT: http://purl.obolibrary.org/obo/NCIT_
  OMIABIS: http://purl.obolibrary.org/obo/OMIABIS_
  OBI: http://purl.obolibrary.org/obo/OBI_

mappings:
  fairgenomes:
    sources:
      - [study.csv~csv]
    s: ega.dataset:$(accession_id)
    po:
      - [a, NCIT:C63536]
      - [OMIABIS:0000006, ega.dataset:$(accession_id)~iri]
      - [OMIABIS:0000037, $(title)]
      - [OMIABIS:0000036, $(description)]
      - [OMIABIS:0000036, $(abstract)]
      - [OBI:0500000, $(study_type)]

Note: The example above be run using an online editor with the following example study.csv.

Mappings Care-SM:

[ ] analyses
[ ] dataset
[ ] experiments
[ ] runs
[ ] samples
[ ] studies

Mappings EJP-RD metadata model:

[ ] analyses
[ ] dataset
[ ] experiments
[ ] runs
[ ] samples
[ ] studies

Mappings Fairgenomes-semantic-model:

[ ] analyses
[ ] dataset
[ ] experiments
[ ] runs
[ ] samples
[ ] studies

wna-se commented 3 months ago

Mapping to CARE-SM turned out to be somewhat challenging. Inquired about support / guidance in CARE-SM/CARE-SM-Implementation #7 Looking for instructions on how to relate the implementation to the CARE-SM data models. Will focus on the EJP-RD metadata model and Fairgenomes-semantic-model to start with and look for a way to transform the EGA model to something compatible with SPARQL.

wna-se commented 3 months ago

This will be added to NBISweden/ejprd-data through #2

NBISweden / ejprd

Map EGA metadata to semantic models #18