Transcript Slide 1

Literature Informatics
Beyond PubMed:
Next Generation
Literature Searching
Carrie Iwema, PhD, MLS
24th August 2011
Growth of PubMed citations from 1986 to 2010
Lu, Database 2011
HSLS, U.Pitt
Information Overload?
HSLS, U.Pitt
GoPubMed: “searching is now sorted”
Ontology-based literature searching
– Medical Subject Headings (MeSH)
• Hierarchical vocabulary for biomedical and health-related topics
– Gene Ontology (GO)
• Controlled vocabulary for molecular biology topics
4 filter categories:
•What
•Who
•Where
•When
Statistics!
Related search engines:
•Go3R
•GoGene
•GoWeb
Developed by Transinsight GmbH
HSLS, U.Pitt
GoPubMed
http://www.gopubmed.org
HSLS, U.Pitt
LigerCat:
“Literature & Genomics Resource Catalog”
•Search articles/journals/genes
•Explore tag clouds
•Craft queries to PubMed
•View Publication History
http://ligercat.ubio.org/
Developed by the Biology of Aging project at MBLWHOI Library
HSLS, U.Pitt
Pubget: “Find papers fast.”
•
•
•
•
The search results ARE the papers—PDFs!
Synched w/home institution journal subscriptions
Customizable “latest issues” journal list
Easily browse your favorite journals
PaperPlane
http://pubget.com/site/help/paper_plane
Developed by Pubget
HSLS, U.Pitt
Pubget
http://pubget.com
HSLS, U.Pitt
eTBLAST: text similarity-based search engine
Uses natural language processing,
keyword weighting, and sentence
alignment to search MEDLINE (and more)
for query-similar text.
Uses:
–
–
–
–
Find an expert
Find a journal
View publication history
Identify implicit keywords
Developed by UT Southwestern Computational Biology Group,
serviced by Virginia Bioinformatics Institute eTBLAST team
HSLS, U.Pitt
eTBLAST
http://etest.vbi.vt.edu/etblast3/
HSLS, U.Pitt
Deja Vu: database of highly similar & duplicate citations
• Offshoot of eTBLAST
• Identifies articles from Medline
exhibiting similar if not identical text
• Plagiarism buster!
Classifications:
– Distinct
– Duplicate
– Erratum
– Sanctioned
– No abstract
– Unverified
Developed by UT Southwestern Computational Biology Group,
serviced by Virginia Bioinformatics Institute eTBLAST team
HSLS, U.Pitt
Deja Vu
http://dejavu.vbi.vt.edu/dejavu
HSLS, U.Pitt
Why They’re Cool…
• GoPubMed— Statistics!
• LigetCat—
Word Clouds!
• Pubget—
PDFs!
• eTBLAST—
Text Similarity!
• Deja Vu—
Plagiarism Buster!
HSLS, U.Pitt
Systems
Year
Major features
RefMed
Quertle
MedlineRanker
MiSearch
Hakia
SemanticMEDLINE
MScanner
eTBLAST
PubFocus
Twease
Clustering results into topics
Anne O’Tate
McSyBi
GoPubMed
ClusterMed
XplorMed
Extracting and displaying semantics and relations
MedEvi
EBIMed
CiteXplore
MEDIE
PubNet
Improving search interface and retrieval experience
iPubMed
PubGet
BabelMeSH
HubMed
askMEDLINE
SLIM
PICO
PubCrawler
2010
2009
2009
2009
2008
2008
2008
2007
2006
2005
Featuring multi-level relevance feedback for ranking
Allowing searches with concept categories
Finding relevant documents through classification
Using implicit feedback for improving ranking
Powered by Hakia’s proprietary semantic search technology
Powered by cognition’s proprietary search technology
Finding relevant documents through classification
Finding documents similar to input text
Sorting by impact factor and citation volume
Query expansion with relevance ranking technique
2008
2007
2005
2004
2001
Clustering by important words, topics, journals, authors, etc.
Clustering by MeSH or UMLS concepts
Clustering by MeSH or GO terms
Clustering by MeSH, title/abstract, author, affiliation, or date
Clustering by extracted keywords from abstracts
2008
2007
2006
2006
2005
Providing textual evidence of semantic relations in output
Displaying proteins, GO annotations, drugs and species
EBI’s tool for integrating biomedical literature and data
Extracting text fragments matching queried semantics
Visualizing literature-derived network of bio-entities
2010
2007
2006
2006
2005
2005
2004
1999
Allow fuzzy search and approximate match
Retrieving results in PDFs
Multi-language search interface
Export data in multiple format; visualization; etc
Converting questions into formulated search as PICO
Slider interface for PubMed searches
Search with patient, intervention, comparison, outcome
Alerting users with new articles based on saved searches
http://www.ncbi.nlm.nih.gov/CBBresearch/Lu/search
/
Lu Z Database 2011
Ranking search results
HSLS, U.Pitt
Lu Z Database 2011
Video Tutorials
• Searching using MESH terms:
http://media.hsls.pitt.edu/media/clres2705/mesh.swf
• Pubmed Clinical Queries:
http://media.hsls.pitt.edu/media/clres2705/scz.swf
• GoPubmed:
http://media.hsls.pitt.edu/media/clres2705/gopubmed.swf
HSLS, U.Pitt
Thanks for your attention.
Good luck searching!
Carrie Iwema, PhD, MLS
Information Specialist in Molecular Biology
Health Sciences Library System
University of Pittsburgh
[email protected]