Binding Workflow Proposal
Automatically generating GO annotations from Binding databases
Intra-species
Binding database: X binds Y, where X and Y are in the same species
Generates two annotations:
- GeneProduct: X
- GO: GO:0005515
- With: Y
- Evidence: IDA
- Col16: Y
- GeneProduct: Y
- GO: GO:0005515
- With: X
- Evidence: IDA
- Col16: X
Inter-species
Binding database: X binds Y, where X and Y are in different species.
Generates two annotations:
- GeneProduct: X
- GO: GO:0005515
- With: Y
- Evidence: IDA
- Col16: ortholog_of(Y)
- GeneProduct: ortholog_of(Y)
- GO: GO:0005515
- With: X
- Evidence: IDA
- Col16: X
Automatically generating Binding database entries from GO annotations
If we have a GO annotation
- GeneProduct: X
- col8: Y
Automatically generate a binding database entry for (X,Y)
Evidence Code
We can "push down" annotations to more specific terms. See http://www.slideshare.net/cmungall/go-pathwayinteractionintegration slide9
This can largely be done outside the workflow here.
One problem is that this is based on PRO which does not cover all species
Evidence Codes
Binding databases have more granular evidence types than GO, see PSI formats. This is not a problem for bindingDB->GO because we can map to the more generic IDA. But it may be a problem for the opposite direction?
Eventually ECO should merge with the evidence types used by PSI formats