Which data from Wikidata would be useful for curation in PathVisio?

[mkutmon]

Let's collect some ideas on what kind of data from Wikidata would be useful when curation biological pathways in PathVisio.
It might be useful for the curation process itself (plugin in PathVisio) or for providing additional info about the curation status of the pathway (curation reports).

[mkutmon]

propose publications for a gene / protein / metabolite

[mkutmon]

propose cellular location of gene/protein
https://upload.wikimedia.org/wikipedia/commons/b/b9/Biomedical_Knowledge_Graph_in_Wikidata.svg
(protein > cell component > cellular component)

[mkutmon]

check for drugs in pathways and see if the interactions with the proper targets are present in the pathway
https://upload.wikimedia.org/wikipedia/commons/b/b9/Biomedical_Knowledge_Graph_in_Wikidata.svg (medication < physically interacts with < protein)

[mkutmon]

identify TextLabels with Disease names and add identifiers as Comments (future: annotation of Disease data node)

[mkutmon]

propose interesting genes for the disease

[mkutmon]

identify TextLabels with BiologicalProcess names and add identifiers as Comments (future: annotation of Pathway data node)

[mkutmon]

identify taxons used in pathway based on gene / protein identifiers
https://upload.wikimedia.org/wikipedia/commons/b/b9/Biomedical_Knowledge_Graph_in_Wikidata.svg
(gene > found in taxon)

[mkutmon]

https://github.com/PathVisio/BH2021/blob/main/queries/multi-species-pathway.rq
https://github.com/PathVisio/BH2021/blob/main/queries/other-species-genes.rq

[ammar257ammar]

I wrote those two queries to get all genes/proteins having more than one taxon assigned to them.

------------------- Genes with more than one taxon --------------------

select ?gene (GROUP_CONCAT(?taxons) AS ?taxon) (COUNT(?taxons) AS ?count)
where{
?gene wdt:P31 wd:Q7187.
?gene wdt:P703 ?taxons.
}
GROUP BY ?gene
HAVING (?count > 1)

------------------- Proteins with more than one taxon --------------------

select ?protein (GROUP_CONCAT(?taxons) AS ?taxon) (COUNT(?taxons) AS ?count)
where{
?protein wdt:P31 wd:Q8054.
?protein wdt:P703 ?taxons.
}
GROUP BY ?protein
HAVING (?count > 1)

---

Protein: 22 results
Gene: 3 results

[mkutmon]

look up ComplexPortal identifiers for existing complexes in WikiPathways

[mkutmon]

still stuck with making the query (missing step = check that all complex proteins are present in the pathway)

SELECT DISTINCT ?wpid ?complex (GROUP_CONCAT(?protein;SEPARATOR=" ") AS ?protein_group) (GROUP_CONCAT(?uniprot;SEPARATOR=" ") AS ?uniprot_group) (count(?uniprot) AS ?count) WHERE {
  ?complex wdt:P31 wd:Q107509287.
  ?complex wdt:P703 wd:Q15978631.
  ?complex wdt:P527 ?protein.
  ?protein wdt:P352 ?uniprot.
  ?pathway wdt:P31 wd:Q4915012.
  ?pathway wdt:P703 wd:Q15978631.
  ?pathway wdt:P2410 ?wpid.
  ?pathway wdt:P527 ?gene.
  ?gene wdt:P688 ?protein.
  }
GROUP BY ?wpid ?complex
HAVING (?count > 1)

[mkutmon]

solution: https://github.com/PathVisio/BH2021/blob/main/queries/identifying-complexes.rq

[mkutmon]

it would be useful to be able to extract literature references from ComplexPortal but they are not in Wikidata yet.

[lubianat]

Added a ticket for that: lubianat/complex_bot#16

[mkutmon]

gene expression data from wikidata integrated with pathways

[marvinm2]

Make sure that GeneProduct datanodes use official HGNC symbols as their labels.

[marvinm2]

select distinct (str(?label) as ?geneProductlabel) (fn:substring(?HGNCsymbol,37) as ?HGNC) where {
?geneProduct a wp:GeneProduct ; rdfs:label ?label ; dcterms:isPartOf ?pathway ; wp:bdbHgncSymbol ?HGNCsymbol .
?pathway a wp:Pathway ; wp:organismName "Homo sapiens"^^xsd:string .
FILTER((str(?label) as ?geneProduct) != (fn:substring(?HGNCsymbol,37) as ?HGNC))
MINUS {?pathway wp:ontologyTag cur:Reactome_Approved}
}

[marvinm2]

This finds all genes with non-HGNC labels in WikiPathways
Next step: identify those pathways they are part of (doesn't work with this query by printing the pathway as output)