Integrative Bioinformatics 2007, Day 1: VINEdb, Hariharaputran et al

Other than where specified, these are my notes from the IB07 Conference, and are in no way expressions of opinion, and any errors are probably just due to my own misunderstanding

Motivation: to create a data warehouse, not only to integrate the diverse data available from different sources but also to introduce a technique that is able to project the underlying data and its complexity in an effective way to the end user. Related works include the Atlas data warehouse. The problem is that the huge data sources are difficult to handle and time consuming, and long lists of data are not sufficient to show what's really going on. The solution is a graphical representation of the facts, and should be easily human readable.

VINEdb: platform-independent. Images created are made at run-time, and there is free web access. THe core is in Java, with MySQL. JSP+Apache for the front-end and Graphviz for drawing. There is a four-layer architecture: source (with data sources), integration (monitor component), data warehouse, and web interface layer.

Here comes my opinion: The talk itself was easy to understand, though a little heavy on the text. I had no problem with the talk itself, just with the fact that they never seemed to get to the meat of it. Even after 20 minutes, I'm simply unsure how their data integration actually works, beyond a data warehouse with accompanying input and output layers. A single architecture cartoon would have worked wonders.  From the screenshots, it doesn't seem as if the front-end is showing something truly new in the GUI design. Perhaps it is a noteworthy new technology, but it is hard to tell.