Annotation data are exported in two separate files each containing different types of functional annotations. Data are exported on a nightly basis in a cumulative file. 


(A) Files named as reviewedGOPOAnnotations.gaf are Gene Ontology (GO) and Plant Ontology (PO) Annotations combined in a single file in GO GAF2.0 format


Full reference from GO site:  
http://geneontology.org/page/go-annotation-file-format-20


Column number:content

1. Database: Source Database of Annotated Object. One of UniProtKB or TAIR or RNACentral.

2. Database Object Identifier :the unique identifier for an object in the Database from column one.  

3. Database Object Symbol : the symbolic name of the object (gene, protein, locus, RNA)  being annotated. Can be a gene product symbol or ORF name. Usually something with biological significance.

4. Qualifier: term that qualifies the relationship between the object and ontology term (e.g. NOT). 

5. Ontology ID: the unique identifier for a GO or PO term

6. Reference ID: the unique identifier for the reference used for annotation. The format is source:identifier. Source is either  PubMed Database (e.g. PMID:23445566) or DOI (e.g. DOI:10.1073/pnas.1713574114)

7. Evidence code: three letter code corresponding to a GO evidence code (one of IDA, IGI, IPI, IMP, IEP).See http://www.geneontology.org/GO.evidence.html for details on evidence codes.

8. Evidence With: Used for some annotations (IPI and IGI) that require supporting information about interacting component (for example a protein binding partner or a genetic supressor). Format is DBname:DBidentifier. More than one entity is allowed. When more than one entity is included a pipe (| )is used to indicate an OR relation, and a comma (,) is used to represent AND relation.

9.  Aspect: Refers to the namespace or ontology aspect. F=GO molecular function, C=GO cellular component, P=GO biological process, S=PO structure, G=PO growth and development stage 

10. Database Object Name: name of the gene, gene product.

11. Database Object Synonym: Additional symbolic names for the gene product. Used to aid searching.

12. Database object type: A description of the object (from column 2) that is being annotated.  May be one of : gene_product, protein, RNA

13. Taxon: unique identifier corresponding to the taxon ID of the gene product being annotated (from column 2).

14. Date: Date on which the annotation was made; format is YYYMMDD.

15. Annotator: The ORCiD of the individual who make the annotation.

16. Annotation Extension: date the annotation was made.


17. Gene Product Form ID:


(B) Files named as otherAnnotations.json are files containing gene name, comment annotations and associations to publications, in JSON format. At a minimum each set will have a gene-publication link, other values may be null.


Example:
{
“publication”: "PMID:16891302",
“locus”: "Q08IT7",
“source”: "UniprotKB",
“symbolicName”: "GmICHG",
“fullName”: "SOYBEAN Isoflavone conjugate-specific beta-glucosidase",
“comments”: []
}


1. publication: either PubMed ID (PMID) or DOI
2. locus: one of UniProtKB id, AGI locus identifier, or RNACentral id
3. source: one of UniProtKB, TAIR, or RNACentral
4. symbolicName: may be null
5. fullName: may be null
6. comments: may be empty list