ECCB 2002 Poster sorted by: Author | Number

Next | Previous poster (in order of the view you have selected)
Title: Mining the literature for enzyme-disease associations
P62
Hofmann, Oliver; Schomburg, Dietmar

o.hofmann@smail.uni-koeln.de
Cologne University Bioinformatics Center

An ever increasing number of diseases and predispositions is being linked to specific enzymes. As the amount of research publications increases and expert knowledge widens new methods to explore and store the available information are needed.

We developed a system that automatically extracts information of enzyme-related diseases from the biological literature and stores this
information in a relational database. The text is lexically filtered and assigned to semantic categories by using MetaMap [1], which maps phrases to the Unified Medical Language System (UMLS, [2]).
A network of associations between enzymes and diseases is built based both on already assigned keywords as well as text and concept similarity using term frequency methods. Enzymes are represented as families and their EC numbers, diseases as Medical Language Subject Headings (MeSH terms). The concept based database allows specific searches on individual diseases or more generic searches (e.g. all links between diseases based on pharmaceutical drugs).
Currently, enzymes from over 600 EC-numbers are connected to about 1.600 different diseases. The network can be displayed as a graph for easier navigation, with additional information like automatic summaries or nearest neighbours being shown interactively.
The approach is useful to shorten the time needed for an overview of a particular field or generate a new hypothesis based on previously undiscovered correlations. The extracted information is added to the enzyme database BRENDA [3].
[1] Aronson AR (2001): Effective mapping of biomedical text to the UMLS Metathesaurus: the MetaMap program. Proc AMIA Symp p17-21.
[2] Humphreys BL, Lindberg DAB, Schoolman HM, Barnett GO (1998): The Unified Medical language System: An informatics research collaboration. J Am Med Inform Assoc. 5(1), p 1-11.
[3] Schomburg I, Chang A, Hofmann O, Ebeling C, Ehrentreich F, Schomburg D (2002): BRENDA: a resource for enzyme data and metabolic information. Trends Biochem Sci 27(1), pp 54-6