Enrich dataset obtained with CREW - ASPIS pipeline (EPIC: INCREASING INFORMATION CONTENT FOR RASAR) - Githubissues

ontox-hu / aspis4j

an api for the ontox4j graph database

0 stars 0 forks source link

Enrich dataset obtained with CREW - ASPIS pipeline (EPIC: INCREASING INFORMATION CONTENT FOR RASAR) #9

Open Maddocent opened 2 years ago

Maddocent commented 2 years ago

To enrich the current version of the CREW-ASPIS-chemicals-dataset (see https://github.com/ontox-hu/ontox-nlp-iris/blob/main/code/aspis_select_workflow.ipynb) we could do all or multiple of the following:

[ ] Connect compound data from CTD to CREW-ASPIS nodes
[ ] Rerun CREW on fulltext
[ ] Refine Pubmed query to include more adversities
[ ] Add in vitro data from ToxCast and other sources
[ ] Add physio-chemical properties
[ ] Run XPaths (pheotype-to-pathway-mapping) pipeline on all phenotypes
[ ] Run entity linker to connect phenotypes to ontologies
[ ] Add compound fingerprints / graphs as node attributes
[ ] Add data from non-mammalian species
[ ] Add more chemicals (whole of Pubmed?)
[ ] Add PMID to relation (COMPOUND) -[r{pmid:pmid_article}]-(PHENO)
[ ] Connect data from cholestasis data extraction project
[ ] Improve CREW (see issue: #10)
[ ] Improve NER (see issue #11)
[ ] Improve Entity linker (see issue #12)

tomlue commented 2 years ago

This data may need to be reformatted to fit the schema that we end up with, but it is a good step after #5