Annotation guidelines for annotating complexes as annotation objects

From GO Wiki
Jump to navigation Jump to search

Annotation guidelines for annotating Complexes as objects

Since its inception, GO has annotated gene products (i.e. proteins or RNA) as the object of annotation. The GOC has recognized the importance of capturing function and process annotations for macromolecular complexes as objects. This document provides guidelines for annotation complexes.

  1. GO Annotations that are shown in IntAct or Pro shouldn't be different from what is exported to GO. This will cause unnecessary data propagation issues and will be hard to track provenance.
  2. Column 2 of GAF should have a stable ID of the macromolecular complex. So far IntAct ID or PRO ID are allowed.
  3. Since there is overlap between GO and IntAct complexes, no need to restate that the Intact complex is_a GO complex. It is okay to annotate that the IntAct Complex to CC terms like mitochondrion or cytoplasm or to say that the IntAct complex is a specific type of a generic GO complex. IPI cannot be used for this anntoation since With column is mandatory for IPI. IDA ? Discuss further?
  4. Do not annotate complexes that have been inferred by chemogenomics. IntAct will use ECO terms to prevent export of such complexes.
  5. If IntAct has Exp evidence for a mouse complex and has inferred the complex for human using ISO, should the human complex get GO annotations in MF and BP? This is fine. In this case, put the Mouse complex_identifier in With/From column.
  6. Can IMP be used an evidence for MF or BP annotation to a complex? If one subunit is deleted and the MF/BP is inferred then the deleted subunit ID should be in With/From column.
  7. If a MF or BP annotation is inferred by curator for a complex (we need example for when this would happen since all complexes have to have a function when it is created in IntAct), then the PMID should be used as reference and instead of GOID, use the IntAct ID in With/From column. This is a new feature. Make a github tracker for this so the checking script allows complex ids in column8.
  8. Regarding X binding, do not annotate heme binding as a function for hemaglobin complex or maltose binding for maltose transport complex and so on.
  9. When human protein is expressed in mouse- need to document this case with examples