Table 2 Summary of the rule model

    Details of classification quality, annotations and classifications are found in Tables 3 and 4


 

a. Annotations, Rules and Classifications

Annotated genes 

within the 23 broad classes of GO biological process

273

Gene probes 

associated with the 273 genes within the 23 broad biological process classes

284

Training examples 

annotations associated with the genes in the 23 broad biological process classes

co-annotationsassociated with the genes in the 23 broad biological process classes

549

444

Rules generated from the training examples

18064

Estimated quality of classifications of unknown genes (cross-validation estimates) 

 

Sensitivity

84%

Specificity

91%

Fraction of classifications that are correct

49% 

 

 

Classifications for unknown (uncharacterized) genes

548

classifications were obtained for 211 of the 213 unknown genes

 

 

 

(Re-)Classifications for training examples

728

True positive classifications519

 

True positive co-classifications ii356

 

False positive classifications219

 

False negative (missing) classifications30

 

For 272 of the 273 training examples at least one correct (re-)classification was obtained

 

 

 

ipairs of two different biological processes annotated to the genes in the dataset

iiclassification of two different biological processes to one gene


 

b. Number of biological processes annotated or classified per gene

Number of biological processes per gene

Annotations for 

training example genes

(Re-)Classifications for

training example genes

Classifications for

unknown genes 

1

105

30

27 

2

100

93

84

3

41

96

59

³ 4

27

54

41