Managers 13Sept06
Present: Michael Ashburner, Judy Blake, Mike Cherry, Rex Chisholm, Jennifer Clark, Midori Harris, David Hill, Eurie Hong, Suzi Lewis, Chris Mungall
Minutes: Eurie
Scientific Advisory Board Meeting Debriefing
We discussed our impressions of take home messages from the SAB meeting. If any of the managers would like a copy of the pink sheets, they can get a hard copy.
Metrics
One of the major messages conveyed by the SAB was the need for metrics. We have developed metrics covering the breadth and depth of the Reference Genome effort and the SAB encouraged defining metrics for ontology development as well as define progress and activities in terms of dollar amounts. Judy mentioned that Peter Good has indiciated that NIH and NHGRI are moving towards measurable contracts. Although some types of financially based metrics may be too extreme, Suzi pointed out that we do need to consider the money aspect.
For ontologies
David mentioned that he has been thinking about how to define metrics for the ontologies. It can be difficult and sometimes not appropriate to define the quality of the ontology using numbers but there are some things you can measure in the ontologies:
- number of terms
- number of relationships
- average relationship per term
- avaerage depth of each tree
Chris already runs a script nightly that measures these statistics on OBO files! Right after the call, he sent out these two URLs: Main GO/OBO stats page and Page of statistics
High numbers don't necessarily mean anything. You could clean up a graph and reduce the numbers but it's still an improvement to the ontologies. These numbers can be used to describe a major change.
There Also need to document the process of decision making.
ACTION ITEM: David and Midori will have a proposal for documentation process and identifying metrics for the January meeting
Other areas
Other suggestions for metrics were keeping track of the number of publications that use GO and keep track of number of public databases that integrate GO terms. All this information could be displayed on a web page on our site. We may need to develop a structure for monthly reports in order to track such information consistently.
Also, we'll need to look into software for tracking bug fixes and feature adding.
ACTION ITEM: Suzi will propose across the board metrics for the group.
Expansion of AmiGO
The advisors did like the idea of expanding AmiGO to be a hub. Chris, Jane, and Eurie have created a smaller group of people consisting of members of the AmiGO working group, the software group, and the production group to shepherd these changes.
Judy mentioned that AgBase has a nice simple entry and basic search. Users can pick any subset of organisms or all organisms to do their query. And suggested the idea of just merging the AmiGO search page with the main GO home page. This is an idea that Jane had proposed as well.
Another considerations for AmimGO is to highlight the work done by the Reference Genome group and displaying the orthology clusters.