BIOLINK 2008: Linking database submissions to primary citations with PubMed Central
- 1. Linking Database Submissions to Primary Citations with PubMed Central Heather Piwowar and Wendy Chapman Department of Biomedical Informatics University of Pittsburgh BioLINK 2008
- 4. These links are important for several reasons
- 8. Sometimes the links are easy to discover
- 13. But the meaning of hyperlinks is ambiguous:
- 14. And often no hyperlinks at all:
- 15. One way to identify links: NLP systems that identify statements of shared data from within full text.
- 16. BUT this requires developing and maintaining a full-text archive!
- 17. What about using PubMed Central?
- 18. Usage? scientists looking for datasets for reuse curators looking for primary citations researchers studying data sharing behaviour
- 19. Goal: Use the simple, full-text query interface of PubMed Central to identify articles with shared datasets
- 20. Method: Gene expression microarray data GEO database
- 21. Method: Open Access articles to train Non-Open access articles to test Gene-expression articles selected by MeSH term query
- 22. Gold Standard: True positives (N=550) Articles with primary citation links from GEO + screening of full-text True negatives (N=165) The rest
- 23. Building the query: Used full-text of open-access cohort Removed words <40 occurrences Unigram bag-of-words vectors Tree and Rule algorithms, a variety of parameters
- 24. (geo OR omnibus) AND microarray AND "gene expression" AND accession NOT (databases OR user OR users OR (public AND accessed) OR (downloaded AND published))
- 25. (geo OR omnibus) AND microarray AND "gene expression" AND accession NOT (databases OR user OR users OR (public AND accessed) OR (downloaded AND published))
- 26. (geo OR omnibus) AND microarray AND "gene expression" AND accession NOT (databases OR user OR users OR (public AND accessed) OR (downloaded AND published))
- 27. (geo OR omnibus) AND microarray AND "gene expression" AND accession NOT (databases OR user OR users OR (public AND accessed) OR (downloaded AND published))
- 28. (geo OR omnibus) AND microarray AND "gene expression" AND accession NOT (databases OR user OR users OR (public AND accessed) OR (downloaded AND published))
- 29. Evaluation Results 40% recall 94% precision, 65% for those not yet linked worse than full-NLP results (~ 89%,83%) slightly better than trivial query (34%,90%)
- 30. Limitations only one datatype database-centric performance so far is rather mediocre…
- 31. Impact? Today’s performance: would increase GEO links by 2.6% by 5.5% annually when all NIH in PMC Double the recall, to 80%: double the numbers above GEO curators added the 40 links identified by this study
- 32. We hope this work inspires future enhancements, and highlights the opportunities for simple full-text queries in PubMed Central given the mandated influx of NIH-funded research reports.
- 33. Thank you Advisor: Dr. Wendy Chapman Funders: NLM and Pitt DBMI Enablers: Everyone who deposits their publications in PubMed Central! My shared data: www.dbmi.pitt.edu/piwowar Share your research data too!