Skip to main content

Table 2 Overall Performance of REGAL on B. napus.

From: Erratum to: genetic algorithm learning as a robust approach to RNA editing site site prediction

 

Known Edited Sites

Total: 258

Known Unedited Sites

Total: 263

 

Predicted Edited Site

True positive

229

False positive

51

Sensitivity: 0.89

Specificity: 0.81

Predicted Unedited Site

False negative

29

True negative

212

PPV: 0.82

Accuracy: 0.85

  1. The performance of REGAL on the B. napus mitochondrial genome is shown here. REGAL was tested on 397 known edited sites and an equivalent number of known unedited sites. The overall accuracy in this genome was 77%, with sensitivity of 83% and specificity of 72%. Of the full set of known edited and unedited sites, 258 known edited sites and 263 known unedited sites were in the 90% credible intervals. Because the numbers of true positives and true negatives are slightly different, PPV as well as specificity are shown.