Table 1.
Accuracy of the PFD-based coding-region identification algorithm on different coding/noncoding subsets. The parameters N1 and N2 are the numbers of coding and noncoding sequences with length greater than n1 and n2, respectively. A DNA segment is declared “coding” if MAXFD > δ0 and “noncoding” otherwise. Accuracy is defined as the average of sensitivity and specificity. The threshold δ0 is set at where sensitivity equals specificity.
| n1 | N1 | n2 | N2 | δ0 | Sensitivity/specificity | |
|---|---|---|---|---|---|---|
| w = 512 | w = 128 | |||||
| 1 | 4125 | 1 | 5993 | 0.1800 | 82.5% | 84.7% |
| 256 | 4067 | 256 | 4186 | 0.1660 | 85.7% | 86.7% |
| 512 | 3756 | 512 | 1948 | 0.1500 | 89.8% | 89.4% |
| 1026 | 2674 | 512 | 1948 | 0.1620 | 92.5% | 91.2% |
| 1026 | 2674 | 1026 | 650 | 0.1320 | 95.4% | 94.4% |