Cite as: eLife 2020;9:e52614 DOI: 10.7554/eLife.52614
Wikidata
- unlimited scope
- structured text
- 750 million statements
- 61 million items
- 12,000 active users
- 100 active computational bots: imports of large structured databases
Wikipedia
- unlimited scope
- mostly free text
FAIR
- Findable
- Accessible
- Interoperable
- Reusable
Data integration
Knowledge integration
query individual databases
stability?
harmonizing
formats
licensing
😊
😟
anyone can publish data
(following guidelines)
centralized resources
difficult
expensive
common data model
good queries
Data integration
Knowledge integration
Wikidata
community of contributors
- domain experts
- bot developers
- queryable SPARQL
👏
👍
😊
heterogeneous knowledge graph
- genes and proteins
- genetic variants
- chemical compounds (e.g. drugs)
- pathways
- diseases
- references
bot automation
retrieve, transform, normalize, upload
Wikidata Integrator (WDI): Jenkins
identifier translation with SPARQL
- limitless scope
- high performance
- high availability
- real-time community editing
integrative queries
crowdsourced curation
Wikidata vs Disease Ontology
- Late 2015: seeded Wikidata with items from Disease Ontology
- 2018: start comparison
-
2030 new cross references GARD + MeSH
- 98.9% correct
-
2030 new cross references GARD + MeSH
BOQA analysis of suspected cases of the disease Congenital Disorder of Deglycosylation (CDDG)
Human Phenotype Ontology
-derived annotations (of 273 extra)
Rephetio algorithm on Wikidata knowledge graph
biomedically-focused subgraph of Wikidata
- 19 node types
- 41 edge types
FAIR
- Findable ✅
- Accessible ✅
- Interoperable ✅
- Reusable ✅