biomed lit mining - Dr. Lars Juhl Jensen - June 12 - IATUL 2007

Dr. Lars Juhl Jensen
EMBL-Heidelberg, Heidelberg, Germany

Biomedical literature mining (and why we really need Open Access)

UPDATE: Presentation (PowerPoint) now online.  ENDUPDATE

MEDLINE
17 million citations

too much to read -> literature mining (get a computer to read them)

but to do that, you need access to the papers

discipline: info retrieval - finding the papers
ad hoc retrieval

MEDLINE - abstracts only
but would like to run on full text

next discipline: entity recognition

need synonyms / mapping lists - manual
plus orthographic variation

ihop
http://www.ihop-net.org/UniPub/iHOP/

discipline: information extraction

formalizing the facts - turning text into databases

Jensen et al Nature Reviews Genetics 2006

new discoveries - text mining

http://arrowsmith.psych.uic.edu/arrowsmith_uic/

mining temporal trends

timeline of buzzwords

integration of text and data

http://string.embl.de/

genotype to phenotype

Korbel et al PLoS Biology 2005 heatmap

UPDATE: I'm almost certain he's referencing

Korbel JO, Doerks T, Jensen LJ, Perez-Iratxeta C, Kaczanowski S, et al. (2005) Systematic Association of Genes to Phenotypes by Genome and Literature Mining. PLoS Biol 3(5): e134 doi:10.1371/journal.pbio.0030134

ENDUPDATE

where are we now?

the tools are there... we need the text

Q: how are researchers using tools?
A: unfortunately many of them aren't aware the tools exist

Q: copyright obstacles - collections of abstracts copyrighted (protection of database) - is this a problem?
full text - could authors prepare a second abstract for literature mining specifically?
A: extraction of facts... isn't really copyright violation
rather than having second abstract, just deposit semantic information and data directly into a database

Q: how does this relate to Biomart?
http://www.biomart.org/
A: they are trying to glue together different data sources