The focus of biomedical text-mining has been on automatic data retrieval, rather than including the user in the discovery process. The web is an inherently user-centered information discovery resource. Interactive literature exploration can take advantage of the inherent links between biomedical publications. In the biomedical field, genes are a common denominator. Genes and proteins serve as basic unit of information.

Biomedical knowledge can be seen as networks of concurrent genes and proteins that include relationships ranging from direct physical interaction to less direct associations with phenotypic observations and pathologies.

iHOPerator is a GreaseMonkey user-script that operates over the iHOP (information Hyper-linked Over Proteins) website. It enhances the gene-centered pages on iHOP by providing a compact, configurable visualization of defining information for each gene. Enabling additional data, such as biochemical pathway diagrams collected from third party resources (KEGG) and displaying it all together in the same browsing context.

It is little better than screen-scraping. Content in HTML was not designed for this and is difficult for GreaseMonkey scripts to parse consistently. HTML may change and require changes in your parser. This could make the scripts brittle and unreliable.
A Semantic representation of this data would mediate this problem.

We need new tools to explore and combine data in meaningful ways. The Semantic Web puts data into a machine-readable format so computers can aggregate data and make inferences about relationships.

(Presentation: 2007 Spring BIOINF JC.pdf 13.6MB)