Enzymes and EC mappings
Plan of Action
Plan for setting up a pipeline for automatically incorporating curated catalytic reactions from Rhea into GO MF. 1 and 2 can happen simultaneously.
- Map existing GO catalytic activity terms to Rhea. Will require some text-ming/matching etc. and lots of manual review. Add Rhea xrefs to GO. Submit missing reactions to Rhea.
- grouping classes won't go into Rhea - will need be handled separately. How do we make axioms for these?
- do Rhea want to incorporate our reactions?
- intern to do this?
- Design pattern for catalytic reaction xps. Create xps, review and add to GO.
- bi-directionality of the reactions won't work in OWL. Did we have a solution to this?
- add missing reactions from Rhea to GO. Not sure if we'll want to add all, but certainly everything with an EC. Rhea is available as BioPax, so we will need to convert to OWL and add reactions names, figure out where they fit in the hierarchy.
- on what basis we will auto-classify - use EC mechanism as a grouping?
- naming classes from Rhea - use KEGG?
- Set up a pipeline for importing new reactions from Rhea to GO (upon Rhea release?):
- pull out new reactions from Rhea
- convert BioPax to OWL
- use reasoner to place new classes in GO MF
- manual check
- add to live GO
- interval between imports
- do we recommend people submit new reactions via Rhea rather than GO?
- do we also have a TG template for new catalytic reactions?
Background on Rhea
In the Rhea project the curators from the SIB are currently working on several aspects, including improving the coverage of known reactions, as well as improving the handling and representation of reactions including polymers and macromolecules, in collaboration with colleagues from ChEBI. The main focus at the current time is really improving the coverage of Rhea (chemically balanced reactions), with a view to promoting its use as a resource for metabolic network reconstruction (e.g. Microme, MetaNetX projects) and as a source of reaction descriptions for UniProt.
Rhea can be used as a source for automation of some of the GO--ChEBI cross-products. It confers the benefit of a more sophisticated representation of chemical roles in reactions than just "input" and "output".
However, it might be more efficient if GO just outsources the whole reaction hierarchy to Rhea, and gets everything from Rhea - text definitions, logical definitions, synonyms etc. The GO editors would give Rhea a block of GO IDs and new reaction term requests would go via Rhea. Either way, we would get the ec2go links via go2rhea and rhea2ec
Current SOPs for handling enzyme updates and new terms
Amelia's Guidelines for New Enzyme Requests
S.O.P for multi-step enzyme reactions