Table 1.

Examples of four types of inconsistent Gene Ontology Annotations; 3 term-related and 1 related to evidence codes

Term inconsistency	Type A—Contradictory description of gene regulation function
	Evidence: As expected, rat SYT1 gave rise to efficient Ca2+-promoted fusion activity
	GO term: negative regulation of vesicle fusion (GO:0031339)
	Inconsistency: The evidence describes that SYT1 has a positive regulation of vesicle fusion which is contradictory to the negative expression in the selected GO term
	Type B—Over-specific GO term selection
	Evidence: Recent studies from this laboratory have provided strong evidence that endogenous GRK2 and GRK6 can regulate the responsiveness of M1 mACh receptor signalling in cultured rat hippocampal neurons
	GO term: negative regulation of G-protein coupled receptor protein signaling pathway (GO:0045744)
	Inconsistency: The provided evidence only support the regulation of G-protein coupled receptor protein singling pathway while its negativity is unknown. Thus, the annotated GO term is over-specific
	Type C—Unsupportive evidence text
	Evidence: In order to characterise the most prominent protein changes that arise in livers from rats fed control or ethanol-containing diets with or without betaine supplementation, cytosolic liver proteins were resolved by 1D PAGE.
	GO term: cytosol (GO:0005829)
	Inconsistency: The evidence text has the mention of a GO concept “cytosol” but does not express the cellular component information of any gene product thus the evidence text does not correctly support the selected GO term
Code inconsistency	Type D—Erroneous selection of experiment type GO evidence code
	Evidence: At 22h after pollination, we found pollen tubes in 37.5% of the ovaries following pollinations by EXPB1 pollen. In contrast, we found no pollen tubes in the ovaries at 22h after pollination when the silks were pollinated by expb1 pollen. Together these data indicate that the expb1 pollen grows more slowly in vivo than the EXPB1 pollen.
	Evidence code: IGI
	GO term: pollen tube growth (GO:0009860)
	Inconsistency: The evidence indicate the annotation is based on allelic variation and the experiment is conducted by comparing the single gene to the alleles of the same gene. Thus, the correct evidence code should be IMP instead of IGI