1. Open access – making the most of biomedical literature mining Lars Juhl Jensen EMBL Heidelberg

2. why open access?

3. why biomedicine?

4. why literature mining?

5. M EDLINE

7. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

8. information retrieval

9. finding the papers

10. if you can’t find them …

11. … they don’t exist!

12. ad hoc retrieval

13. users-specified query

14. “yeast AND cell cycle”

15. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

18. M EDLINE

19. abstracts

20. complete papers

23. tricks

24. stemming

25. yeast / yeasts

26. synonyms

27. yeast / S. cerevisiae

28. dynamic query expansion

29. next logical step

30. ontologies

31. annotation

32. Cdc28  yeast gene

33. Cdc28  cell cycle

34. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

35. “yeast AND cell cycle”

36. entity recognition

37. identifying the substance(s)

38. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

39. if you can’t find them …

40. … they don’t exist!

44. abstracts

45. M EDLINE

46. tricks

47. good synonyms list

48. manual curation

49. orthographic variation

50. CDC28

51. Cdc28p

52. disambiguation

53. hairy

54. SDS

55. Cdc2

56. information extraction

57. formalizing the facts

58. co-mentioning

59. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

61. NLP Natural Language Processing

62. Mitotic cyclin (Clb2)-bound Cdc28 (Cdk1 homolog) directly phosphorylated Swe1 and this modification served as a priming step to promote subsequent Cdc5-dependent Swe1 hyperphosphorylation and degradation

63. 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]

65. new discoveries

66. text mining

69. temporal trends

71. buzzwords

73. grant applications

74. global correlations

75. 327983 3592 RegulatesRegulated P < 9  10 -9

76. transcriptional networks

77. 112744 3704 PhosphorylatesPhosphorylated P < 2  10 -7

78. signal cascades

79. 810747 3625 ExpressionPhosphorylation P < 5  10 -4

81. integration of text and data

86. network mining

87. linking genes to diseases

92. multifactorial diseases

93. genotype to phenotype

97. where are we now?

99. abstracts

100. complete papers

101. restricted access

102. open access

103. the tools are there

104. now we need the text!

105. Acknowledgments Jasmin Saric Rossitza Ouzounova Michael Kuhn Isabel Rojas Miguel Andrade Peer Bork