Europe PubMed Central and Linked Data
4 likes747 views
On Europe PubMed Central, we extract identifies (e.g., accession numbers, data DOIs) in scientific articles. Recently, we started publishing mined identifiers on Linked Data Platform to improve the connectivity of our mined data.
Europe PubMed Central and Linked Data
- 1. Biohackathon 2015 Europe PubMed Central and Linked Data Jee-Hyub Kim 0000-0002-0359-2887 Nagasaki 13 Sep 2015
- 2. Contents Europe PubMed Central Linking Literature Mining Identifiers Publishing Mined Identifiers on RDF Web Annotation Data Model Use Case for Database Curation
- 3. Europe PubMed Central Europe PMC is a literature database Abstracts: 30 million PubMed, Agricola and patent records, updated daily Full text articles: over 3 million full text articles, of which over 900,000 are free to read and reuse, updated daily
- 4. Services in Europe PMC RESTful web service: http://europepmc.org/RestfulWebService Text-mined terms, metadata, full text ORCID article claiming tool Embassy Cloud for 3rd party contents providers BioJS literature module: http://biojs.io/d/biojs-vis- pmccitation RSS
- 5. Linking Literature Europe PMC provides various types of linking methods By external links: to any URL (e.g., database, Wikipedia, press release, etc.) By text mining Biological entities Identifiers (e.g., accession numbers) By ORCID (article claims) 24 external links providers, 1 ORCID, 9 cross-reference DBs, 20 DB identifiers, 6 named entity types
- 6. Linking Examples To By Relation REST API Wikipedia Provider Mention labsLinks Publons Provider Review labsLinks UniProt Curator Citation databaseLinks ORCID Provider Author search EFO Named entity tagger Recognition textMinedTerms PDB Accession number tagger Mention textMinedTerms
- 7. Mining Identifiers in Free Text Motivation Started for cross-linking with EBI databases Data citation, impact analysis Now, moving for linked data We use patterns from identifiers.org and link back to it. A IE problem: ID matching + NER for resource names Some ambiguities PDB: 4min OMIM and ERC funding id: both 6-digit numbers Resource name variations: UniProt, Swiss-Prot, etc.
- 8. Mentioned in Europe PMC articles Identifiers in Literature Databases ENA, PDB, ArrayExpress, UniProt, RefSNP, OMIM, PFam, RefSeq, Ensembl, InterPro, Bioproject, Biosample, EMDB, PXD, EGA, TreeFam Funding resources European Research Council Ontologies GO, UniProt, EFO, ChEBI, NCBI Taxonomy, UMLS Clinical Trials NCT, EudraCT Digital Repositories (Dryad, figshare, etc.) Data DOI
- 9. Identifiers in Different Resources Articles (978,605) Patents 2014 (266,192) Wiki pages (15,346,290) db # articles db # patents db # pages ena/genbank/ ddbj 23,295 ena/genbank/ ddbj 4,074 pdb 4,265 pdb 15,544 uniprot 1,387 omim 2,226 nct 13,006 pdb 1,093 uniprot 1,712 refsnp 10,168 refseq 1,002 refseq 1,643 refseq 6,551 refsnp 322 ensembl 1,402 omim 5,093 omim 254 go 1,351 uniprot 2,865 pfam 115 pfam 582 go 1,900 ensembl 97 interpro 560 arrayexpress 1,832 interpro 46 ena/genbank/ ddbj 396
- 10. Publishing Identifiers on RDF Goals More connectivity More provenance for each linking PMCID, sentence, section label, etc. Links to share and comment (e.g., hypothes.is) Challenges: How to model? Web Annotation Data Model. dealing with nearly a billion annotations generated automatically in a large scale
- 11. Web Annotation Data Model Built on the top on RDF Annotations as resources To provide a standard description mechanism for sharing annotations between systems For more general purpose use Not only for text mining For example, YouTube video comments (by people), image annotation, etc. W3C Working Draft
- 12. Core Annotation Framework Typically an Annotation has a single Body, which is the comment or other descriptive resource, and a single Target that the Body is somehow "about". The Body provides the information which is annotating the Target. This "aboutness" may be further clarified or extended to notions such as classifying or identifying.
- 14. Text-Mining RDF Service Running on EBI RDF Platform Stores 1,563,241,810 triples text-mined from 400,746 Open Access articles in Europe PubMed Central. Provides for each article, all the annotations linking to ontologies/databases with contexts: sentences section information
- 15. Use Case for Database Curation Given an database identifier, provides sentence-level information for database curation. Show all the articles where a PDB accession number 3NSS is mentioned. Show all the annotations with each its label in PMC3382907. Show all the articles where inflammatory bowel disease (C0021390) is mentioned. http://wwwdev.ebi.ac.uk/rdf/services/textmining/sparql
- 17. Plans for BioHackathon 2015 Integration with other SPAQL endpoints Interoperability with other formats used in text-mining community e.g., BioC, UIMA Produce more links on RDF
- 18. References Europe PMC Consortium. Europe PMC: a full-text literature database for the life sciences and platform for innovation. Nucleic Acids Res. 2015 Jan;43(Database issue) D1042-8. doi:10.1093/nar/gku1061. PMID: 25378340; PMCID: PMC4383902. Kafkas , Kim JH, McEntyre JR. Database citation in full text biomedical articles. PLoS One. 2013;8(5) e63184. doi:10.1371/journal.pone.0063184. PMID: 23734176; PMCID: PMC3667078. Juty N, Le Nov竪re N, Laibe C. Identifiers.org and MIRIAM Registry: community resources to provide persistent identification. Nucleic Acids Res. 2012 Jan;40(Database issue) D580-6. doi:10.1093 /nar/gkr1097. PMID: 22140103; PMCID: PMC3245029.