YeastHub - the semantic web at work
The current issue of Bioinformatics describes the YeastHub database , created by the Gerstein lab.
Yeast is a well sampled and organism with many high throughput data sets to integrate and several established and curated resources such as MIPS, SGD and what used to be YPD.
The new YeastHub data base connects available data using semantic web technologies such as RSS, RDF and a relational to RDF mapping. The set of technologies has the potential to solve many of the small problems one has do deal with when integrating data across many sources.
Eagerly, I tested the data base but was a little let down: While I see the obvious benefits, many small problems appear, which taken together, make the system not really very helpful to the average bioinformatics user, let alone a biologist in the the yeast community.
The lack of descriptions of formats of the data sources makes it tough to create queries and I did not really get past trivial results. I also dearly missed capabilities to browse the data sets. However, it appears as if these technologies will become more and more important and the old tab separated tables hopefully disappear one fine day.
Yeast is a well sampled and organism with many high throughput data sets to integrate and several established and curated resources such as MIPS, SGD and what used to be YPD.
The new YeastHub data base connects available data using semantic web technologies such as RSS, RDF and a relational to RDF mapping. The set of technologies has the potential to solve many of the small problems one has do deal with when integrating data across many sources.
Eagerly, I tested the data base but was a little let down: While I see the obvious benefits, many small problems appear, which taken together, make the system not really very helpful to the average bioinformatics user, let alone a biologist in the the yeast community.
The lack of descriptions of formats of the data sources makes it tough to create queries and I did not really get past trivial results. I also dearly missed capabilities to browse the data sets. However, it appears as if these technologies will become more and more important and the old tab separated tables hopefully disappear one fine day.
spitshine - 2005-06-21 10:57
ISMB YeastHub talk notes
For example, based on the talk, YeastHub uses wrappers to convert existing data sources to RDF and stores those in a Sesame database. I don't think this is any different from a data wearhouse (i.e. write wrappers, store in a relational database).
The technologies are around for quite some time already and I don't think that the established data bases did solve the problem in a open, easy to add manner but rather created monolithic structures.
Technically, the solutions is not all that different really, I agree.