castanyes blaves

Towards a cyberinfrastructure for the biological sciences: progress, visions and challenges

Lincoln D. Stein

• A lot of these concepts are present in day-to-day conversations here at the WTGC.
  
• The "community annotation hubs".
  
• An example of community annotation: the approach taken by Reactome is in my opinion a very clever one: active attract experts on the field, and manually curate the information of their expertise.
• An example of a large community for manual annotation: the Havana group at the Sanger. Lots of people nowadays (with a big push from the ENCODE consortium).
• There still are some gaps or disconnections between sources of knowledge and centralized databases. For example, there is a lot of gene-by-gene work done that could be easily put together in a Reactome-style way. Many of these cases are something like: "I've done my PhD on the characterization of gene XYZ and here are the corrections to the annotation in your online database". A lot of this stuff is still disconnected at the moment of creation, and when it gets published and the investigator moves forward, the momentum gets lost. Then an extra manual curation/text mining effort is needed a few years down the road.