Ontology meeting 2011-12-21

From GO Wiki
Revision as of 14:21, 1 July 2014 by Gail (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

PRESENT

  • Paul Thomas
  • Rebecca Foulger
  • Jane Lomax
  • David Hill
  • Tanya Berardini


REPRESENTING REGULATION IN GO (particularly signaling)

Signaling pathways are essentially a series of regulation steps, and we need a way to better represent this in GO.

  • Paul Thomas presented a preliminary proposal of representing regulatory events in GO better (with a focus on signaling pathways).
  • Defining pathways as a set of MFs: need to say what the functions are regulating.
  • Way of representing what a molecule DOES to another molecule (if possible, connected to the mechanism by which it does it; regulation at the molecular level).
  • This brings in the idea of having 'roles' in GO, to give two branches of the function ontology. E.g:
 molecular role
 --receptor activity
 --protein-specific activity regulator
 ----biochemical_activity_regulator
 -----biochemical_activity_regulator_by_protein_modification
 -------protein kinase activity
 has Target Effect => some regulation
 has Target Protein => some protein


Using MAPK cascade as an example

  • A MAPK cascade would consist of at least a MAP3K, a MAP2K and a MAPK.
  • Class = MAPK cascade. You would have variations of class, depending on what's in the cascade, but it would have at least:
 MAP4K (0 or more)
   |
   v
 MAP3K (at least 1)
   |
   v
 MAP2K (at least 1)
   |
   v
 MAPK  (at least 1)
  • MAP3K is part of the MAPK cascade, and its role in the cascade is to phosphorylate a MAP2K.
  • See PMID 20811974 for a reference (The MAP Kinase Signaling Cascades: A System of Hundreds of Components)


Potential Issues

  • Some cascades will have hundreds of variations. For example PI3K activates AKT. AKT is a Ser/Thr kinase with HUGE numbers of downstream targets. Therefore, the variations on the class would be large.
  • Need to capture:
  1. Components that signal TO PI3K (E.g IRS1)
  2. Components that signal FROM PI3K (E.g. TOR)
  3. Protein X and PI3K signal in the same pathways, but it's not known whether PI3K is upstream or downstream



How to proceed

  • Keep adding in the signaling pathways VIA x cascade terms, using HAS_PART relationships.
  • Connect MF and BP terms as much as possible (E.g. PI3K signaling HAS_PART PI3K activity).
  • Carry on better defining the limits of the intracellular cascades
  • Paul to talk to Chris M. about his proposal, and to send it round the editors list.