Friday, May 28, 10am, 2405 Siebel Center

“BeeSpace: An Interactive System for Functional Analysis Using Semantic Indexing of Biological Literature”

Dr. Bruce Schatz, Department of Medical Information Science, University of Illinois at Urbana-Champaign (http://www.beespace.uiuc.edu)

BeeSpace is the flagship bioinformatics project in the NSF Frontiers of Integrative Biological Research program.  Supporting new computer science research, we have developed an interactive system for functional analysis.

This uses language processing for information extraction on biological literature, particularly for the new paradigm of automatic extraction on special collections.  Semantic indexes are built to support conceptual navigation of gene function.  Generating these indexes requires extensive computation on parallel computers and cloud clusters, including being the largest internal user of the experimental cloud cluster.  The special collections can be selected via realtime clustering of document collections or via steerable clustering with specified topics.  The navigation can be supported via document summarization of gene functions, using an entity recognition variant with biology instead of newspaper entities.  Direct functional analysis is supported by computing the concept lists of the most discriminating terms within the collection corresponding to a list of genes.  All the services can be interactively combined during a session to locate a region of the information space containing the most useful functional information.  This is a practical bioinformatics application of computer science research into natural language processing.
(Slides from this talk are available here )

Back