WormBase-Caltech Weekly Calls

2012 Meetings

Grant Renewal

Like to have rough draft by end of July
Including things we are doing now that are obvious
30 pages total
10 pages for Caltech, likely
Focus on: Gene function, transcriptional networks, pathways
How much have we curated? What is our rate of curation?
What is the backlog? How much left to curate?
Not caught up on RNAi and allele curation by end of this term
Logical extensions to GO (we will discuss next week)
How are we going to measure quality? Metrics?
- Coverage?
- Accuracy?
- What is a/the gold standard? These things aren't generally well defined
Other species
- Automated triage for papers
- Use nematode Textpresso to estimate what's out there?

LEGO (Logical extensions to GO)

Try to use GO annotations to make more expressive statements about gene function, pathways and processes
Prior to this, curators took gene or gene product to describe subcellular localization, molecular function, and processes the product involved in
Looking to create much bigger picture of the biology
LEGO annotations: adding info to typical GO annotation, such as regulation targets
- TF has sequence-specific DNA binding activity; now add what the targets are
- If TF in nucleus, what cell types?
- Now we could start to create pathways and describe processes in more detail
Being done with OWL, web ontology language
Example, the dauer pathway; do more examples to see where the holes are
8 examples from C. elegans: involving cell types, signal transduction pathways, etc.
Build much broader picture of the biology of the worm
Filling in annotation gaps or inconsistencies
Start to build annotation models
Can we integrate with Wikipathways?
Community annotations?

Protein-to-GO

EBI's protein GO annotation system
Data written to UniProt
We agreed that we could clean up our GO annotation files and try this out
Big picture: many groups may buy in to using Protein-to-GO, but will not become our only curation tool
Can we use the OA to modify or create extensions to GO or Protein-to-GO?
What would we have to do to make the OA better than/comparable to Protein-to-GO?
- We need a drop down for the relations
- We need an ontology domain to match to the relations
One benefit/intention of centralized curation system was to have a central quality control system in place
Don't need a central/unified curation system to have a central database, but may help reduce some effort/work
Protein-to-GO has a lot of real time error checking, which is very helpful
Arguably, common/central curation tool and database could speed up curation across different databases
Wormbase can push ahead with a pilot test