Mining text and data on chemicals
Download as PPT, PDF0 likes247 views
This document summarizes Lars Juhl Jensen's work on text mining of chemicals and medical data. It discusses three parts: (1) using text mining to identify relevant papers on specific topics from the exponentially growing literature, (2) integrating data from multiple sources on chemicals, proteins, genomes and their interactions, and (3) extracting information from unstructured patient records written by doctors. The goal is to develop natural language processing methods to formalize extracted facts and relationships to aid in tasks like pharmacovigilance and disease comorbidity analysis.
![Gene and protein names
Cue words for entity recognition
Verbs for relation extraction
[nxexpr The expression of
[nxgene the cytochrome genes
[nxpg CYC1 and CYC7]]]
is controlled by
[nxpg HAP1]](https://image.slidesharecdn.com/jen11talk21-120307084828-phpapp02/85/Mining-text-and-data-on-chemicals-55-320.jpg)
Mining text and data on chemicals
- 1. Mining text and data on chemicals Lars Juhl Jensen
- 2. three parts
- 3. text mining
- 10. some things are constant
- 12. ~45 seconds per paper
- 14. find the relevant papers
- 15. still too much to read
- 16. computer
- 17. as smart as a dog
- 18. teach it specific tricks
- 23. small molecules
- 24. proteins
- 25. diseases
- 27. synonyms
- 29. “black list”
- 31. Reflect
- 33. browser add-on
- 34. Pafilis, O’Donoghue, Jensen et al., Nature Biotechnology, 2009 O’Donoghue et al., Journal of Web Semantics, 2010
- 35. Firefox
- 37. Google Chrome
- 38. Safari
- 39. Utopia Documents
- 40. web services
- 41. collaboration
- 45. SciVerse
- 53. co-mentioning
- 55. Gene and protein names Cue words for entity recognition Verbs for relation extraction [nxexpr The expression of [nxgene the cytochrome genes [nxpg CYC1 and CYC7]]] is controlled by [nxpg HAP1]
- 57. STITCH
- 58. Kuhn et al., Nucleic Acids Research, 2012
- 61. 1100+ genomes
- 63. physical binding
- 64. chemical–protein
- 65. protein–protein
- 68. drug targets
- 69. complexes
- 70. pathways
- 71. Letunic & Bork, Trends in Biochemical Sciences, 2008
- 72. text mining
- 73. co-mentioning
- 77. many data types
- 78. many databases
- 81. variable quality
- 82. not comparable
- 83. spread over many genomes
- 84. quality scores
- 85. von Mering et al., Nucleic Acids Research, 2005
- 86. calibrate vs. gold standard
- 87. von Mering et al., Nucleic Acids Research, 2005
- 92. a hard problem
- 93. in Danish
- 94. by busy doctors
- 96. no lexicon
- 97. acronyms
- 98. typos
- 99. delusions
- 102. Negation F20 F200 Family
- 103. medication
- 104. adverse drug events
- 105. diagnoses
- 106. pharmacovigilance
- 108. Roque et al., PLoS Computational Biology, 2011
- 109. disease comorbidity
- 110. Roque et al., PLoS Computational Biology, 2011
- 111. DNA sequencing
- 112. genotype
- 113. phenotype
- 114. Acknowledgments Reflect STITCH EPJ-mining Sune Frankild Michael Kuhn Francisco S Roque Heiko Horn Damian Szklarczyk Peter B Jensen Evangelos Pafilis Andrea Robert Eriksson Juan-Carlos Silla-Castro Franceschini Henriette Schmock Michael Kuhn Milan Simonovic Marlene Dalgaard Reinhardt Schneider Alexander Roth Massimo Andreatta Sean O’Donoghue Pablo Minguez Thomas Hansen Tobias Doerks Karen Søeby Manuel Stark Søren Bredkjær Christian von Anders Juul Mering Thomas Werge Peer Bork Søren Brunak
- 115. larsjuhljensen