Table 2.
Accuracy versus Signal Type
| Programs | Signal type | |||
| start codon (195) | acceptor site (753) | donor site (753) | stop codon (195) | |
| FGENES | 0.67 | 0.80 | 0.85 | 0.75 |
| (0.63) | (0.77) | (0.82) | (0.72) | |
| GeneMark.hmm | 0.46 | 0.81 | 0.82 | 0.57 |
| (0.60) | (0.75) | (0.78) | (0.64) | |
| Genie | 0.56 | 0.77 | 0.78 | 0.72 |
| (0.57) | (0.82) | (0.83) | (0.73) | |
| Genscan | 0.61 | 0.87 | 0.90 | 0.76 |
| (0.78) | (0.80) | (0.84) | (0.86) | |
| HMMgene | 0.75 | 0.81 | 0.83 | 0.78 |
| (0.78) | (0.85) | (0.87) | (0.81) | |
| Morgan | 0.43 | 0.66 | 0.65 | 0.39 |
| (0.43) | (0.57) | (0.56) | (0.39) | |
| MZEF | — | 0.59 | 0.66 | — |
| (0.65) | (0.73) | |||
-
For each program, the proportion of actual signals identified correctly (the upper number) and the proportion of predicted signals that are correct (the lower number) are averaged over all signals belonging to a particular type. The number in parenthesis in the header of each column represents the number of signals of each type in the HMR195 dataset.
