Table 1.
Prediction results with the 2880 protein dataset using the ANN
Approach | Coverage (%) | Probability (%) | Accuracy (%) | MCC | ||
---|---|---|---|---|---|---|
Qo(H) | Qo(Nh) | Qp(H) | Qp(Nh) | |||
AA | 58.4 | 78.5 | 68.1 | 63.7 | 65.5 | 0.31 |
MSA | 66.7 | 67.3 | 66.7 | 53.4 | 67.0 | 0.34 |
AA + ACC_O | 60.9 | 72.0 | 68.5 | 65.0 | 66.4 | 0.33 |
AA + SS_O | 67.9 | 74.4 | 72.8 | 70.1 | 71.2 | 0.43 |
Seq–Str network (SS_O) | 58.5 | 80.7 | 75.1 | 66.2 | 69.6 | 0.40 |
Seq–Str network (SS_P) | 56.5 | 77.5 | 71.5 | 64.2 | 67.1 | 0.37 |
AA: amino acid sequence; MSA: multiple sequence alignment; ACC_O: observed accessibility (DSSP); SS_O: secondary structure observed (DSSP); seq: sequence; str: structure; SS_P: secondary structure predicted (PSIPRED).