Skip to main content

Table 4 Performance of N-signal vs N-signal-free protein binary classification

From: Plus ça change – evolutionary sequence divergence predicts protein subcellular localization signals

 

Mean accuracy

Mean AUC

Mean MCC

J48

72.49±3.30

0.68±0.09

0.40±0.09

- (randomized)

65.85±0.66

0.50±0.01

0.00±0.03

SVM

74.64±2.38

0.68±0.03

0.40±0.06

- (randomized)

66.19±0.09

0.50±0.00

0.00±0.00

The majority class fraction

65.98%

N/A

N/A

  1. Three classification performance measures when using only divergence features are shown for the discrimination of N-signal containing and N-signal-free proteins (yeast curated ortholog sets). AUC denotes the area under the ROC curves. (randomized) indicates the values obtained with the localization class labels randomly shuffled 100 times. For each measure the average and standard deviation is shown over the 5 folds of the cross-validation, or 500 (5 × 100 trials) folds in the case of the randomized data.