Annotation Conf. Call, October 25, 2011

From GO Wiki
Jump to: navigation, search


  • Webex invite:

  • Status of groups with regards regularly integrating annotations from PAINT, GOC, UniProt etc? (Emily)
  • Protein binding. We'd like to propose that the child terms of protein binding be limited, currently we have terms that describe the binding of:
  1. binding of specific gene products (e.g. GO:0042988, X11-like protein binding, GO:0051528 NFAT3 protein binding)
  2. binding of protein families (e.g. GO:0010725 TBP-class protein binding)
  3. dependence for a binding activity (e.g. ATP-dependent protein binding, copper-dependent protein binding)
  4. protein domain, or protein region binding (e.g. MADS box domain binding and nuclear localization sequence binding, protein C-terminus binding)
  5. functional consequences of protein binding (e.g. GO:0043495 protein anchor, GO:0032947 protein complex scaffold, GO:0030674 protein binding, bridging, GO:0006993 sterol regulatory element binding protein import into nucleus involved in sterol depletion response )
  6. State of protein bound (e.g. denatured protein binding, misfolded protein binding, preprotein binding)
  7. Binding of proteins within a complex (e.g. GO:0032403 protein complex binding)

Where an interaction can be equally represented by >1 of these descriptive protein binding terms, how do curators decide which one(s) to annotate, and if a curation decision needs to be made, how can a consistent annotation set be created?

Do we need to prune back some of these child terms or protein binding , where we might be venturing into domains best covered by external resources

As a first step, could we introduce protein families (PRO/UniProt), to be able to broadly categorise the identity of the bound identifier (and loose the gene specific binding terms)? Suggestions of resources to use? Currently PRO doesn't seem to have many terms at this level. See SF:3409611 and SF:3423447 for examples.

NB: There will be a small number of specific protein binding terms that we will probably need to retain because they are a key part of a larger processes e.g. signalling, and some are has_part to the receptor activity e.g. interleukin-12 receptor has_part interleukin-12 binding [Jane & Emily]