Table 1.
SVM Peformance in Distinguishing Real From Pseudo Exons
|
Flanks
|
Splice sites
|
|||||
|---|---|---|---|---|---|---|
| US | DS | 3′ | 5′ | Exon body | ROC | Specificitya |
| CVb | 0.609 | 0.484 | ||||
| + | - | - | - | - | 0.791 | 0.638 |
| - | + | - | - | - | 0.784 | 0.618 |
| + | + | - | - | - | 0.855 | 0.695 |
| - | - | + | - | - | 0.823 | 0.672 |
| - | - | - | + | - | 0.837 | 0.698 |
| - | - | + | + | - | 0.907 | 0.777 |
| + | + | + | + | - | 0.932 | 0.825 |
| - | - | - | - | + | 0.946 | 0.841 |
| + | + | - | - | + | 0.984 | 0.956 |
| - | - | + | + | + | 0.987 | 0.964 |
| + | + | + | + | + | 0.991 | 0.976 |
Performances are indexed by ROC values and specificity. Each row is an SVM test. ROC values were measured in untouched sets of ∼2200 real and ∼2300 pseudo exons. The first five columns indicate the components used by SVM. US, upstream; DS, downstream; TP, true positive; FN, false negative; FP, false positive; SE, sensitivity; SP, specificity.
Specificity = TP/(TP + FP) at a sensitivity (SE = TP/(TP + FN)) of 0.90
The SVM classified on the basis of the acceptor and the donor consensus values