Give GO annotation a go - Part II
The following is the second in a series of posts about how experiment based GO annotations are assigned to genes in TAIR.
The manual process of assigning experimental GO annotations to genes starts with a prioritized list of papers describing genes for which no or limited GO information exists in TAIR (see our December 2018 blog post for more information). A TAIR curator reads a paper and determines the most appropriate GO or Plant Ontology (PO) term or terms depending on the area of biology described in the paper (cellular component, molecular function, etc…) along with the experimental method which is captured in the evidence code. In the case of protein-protein interaction results or multi-gene mutants, the additional gene IDs are provided as supporting evidence.
When choosing a GO term we try to select the most specific term possible. Sites such as AmiGO 2 (http://amigo.geneontology.org/amigo) and QuickGO (https://www.ebi.ac.uk/QuickGO/). These websites, which are updated on a regular basis, allow keyword searches and display a set of GO terms, their definitions and relationships. These relationships are displayed in a hierarchical tree view of all related GO terms which aids in identifying the most accurate and specific term.
Image cropped from QuickGO
When assigning a GO term, it needs to be as specific as possible. When we look for a specific GO term in the various websites, but cannot find anything that closely resembles it, we will create a new GO term and get it added to the list with all possible GO terms.
Finally, GO terms are only assigned based on supporting data in the paper. For example, if experimental data only indicate that the gene product binds to DNA then the GO term can only reflect this and nothing further even though the gene is commonly known as a DNA binding transcription factor. We will go over the Evidence component of a go annotation in the next part (III) of our GO blog series.
Hopefully this is helpful and will make you feel like contributing to the GO annotation effort by submitting your paper’s GO info to TOAST (toast.arabidopsis.org).