- From: John Wilbanks <wilbanks@w3.org>
- Date: Tue, 17 Aug 2004 13:57:31 -0400
- To: public-semweb-lifesci@w3.org
All, Here is a use case derived from a conversation I had recently with a computational biologist. 1. I am a researcher with a killer set of taqman data - and I want to share it with the world. 2. I am busy, and perhaps don't understand the data issues inherent in GO. 3. I overlay onto GO assignments for the gene. 4. However, what I've really found is a complex transcriptional region, and what GO has looked at is coding regions with clear functions to annotate. 5. My taqman data is looking at a different exon which is proximal to GO annotation. 6. I release my data with the GO annotation, but its utility is comprimised, as the annotation does not apply to my data My question is: what are the best ways to go after this type of problem, which in many ways is derived from the conflicting perceptions of what "gene" means? Should the W3C be thinking about standards for overlaying ontologies onto raw data, with those standards determined for given entity classes or ontological classes? Should we simply look at making sure our recommended technologies *support* the scientifically accurate overlay of ontological classification onto binary data? The central point is that imprecise usage of ontologies for functional annotation can actively undermine the utility of the information... Please discuss! jtw -- John Wilbanks W3C Fellow Semantic Web - Life Sciences http://www.w3.org/People/all#wilbanks wilbanks@w3.org 617-253-5845 (direct) 617-838-6333 (mobile; best voicemail #) --
Received on Tuesday, 17 August 2004 17:56:59 UTC