Table 3.
Validation dataset | Class | Sensitivitya | Specificityb | LOOCV error rate |
---|---|---|---|---|
Affy FF | Mut | 0.852 = 23/27 | 0.778 = 21/27 | 19 % |
NanoS_FFPE | Mut | 0.704 = 19/27 | 0.741 = 20/27 | 28 % |
Affy_FFPE | Mut | 0.519 = 14/27 | 0.889 = 24/27 | 30 % |
Note: Samples were classified as either KRAS/BRAF mutant (Mut, n = 27) or KRAS/BRAF wild-type (WT, n = 27). Class prediction was performed using the Affy_FF, Affy_FFPE and NanoS_FFPE samples sets
a = number of predicted mutants divided by number of true mutants
b = number of predicted WT divided by number of true WT