Table 2 |
||
|
Transitive annotation-based scale (TABS): a qualitative distance scale for the assessment of annotation reproducibility in genome projects |
||
|
Score |
Description |
Comment |
|
|
||
|
7 |
False positive |
Original annotation predicts function without any supporting evidence |
|
6 |
Over-prediction |
Original annotation predicts a specific biochemical function without sufficient supporting evidence |
|
5 |
Domain error |
Original annotation overlooks different domain structure of query and reference proteins |
|
4 |
False negative |
Original annotation does not provide predicted function although there is sufficient evidence to characterize the query protein |
|
3 |
Under-prediction |
Original annotation predicts a nonspecific biochemical function although a more detailed prediction could have been made |
|
2 |
Undefined source |
Original annotation contains undefined terms, non-homology based predictions, and so on |
|
1 |
Typographical error |
Original annotation contains typographical errors that may be propagated in the database |
|
0 |
Total agreement |
Original annotation is correct, but annotations may be only semantically (but not computationally) identical |
|
|
||
|
Column names and explanations: Score, the score between two assignments; Description, a description of the potential disagreement between two projects; Comment, explanatory comments for ranking/scores. We consider scores of 0-3 as relatively benign compared to scores of 4-7, as the latter have a much more significant impact on genome sequence and database quality. |
||
|
Ouzounis and Karp Genome Biology 2002 3:comment2001.1 doi:10.1186/gb-2002-3-2-comment2001 |
||