Skip to main content

Table 4 Class-specific prediction accuracy for various prediction algorithms

From: Ensemble-based prediction of RNA secondary structures

 

ALL

ASE

CRW

PDB

RFA

SPR

SRP

TMR

n

2511

386

411

311

257

526

350

269

Testset contribution

0.8

0.83

0.79

0.76

0.78

0.78

0.80

0.87

Mean sequence length

332

959

75

129

116

77

226

362

BL-FR*

0.703

0.606 (0.592, 0.620)

0.613 (0.590, 0.637)

0.900 (0.878, 0.920)

0.674 (0.633, 0.713)

0.780 (0.761, 0.800)

0.734 (0.712, 0.755)

0.589 (0.569, 0.607)

BL*

0.688

0.604 (0.589, 0.618)

0.583 (0.561, 0.603)

0.894 (0.871, 0.915)

0.667 (0.627, 0.704)

0.763 (0.742, 0.782)

0.717 (0.693, 0.738)

0.568 (0.550, 0.587)

CG*

0.676

0.601 (0.588, 0.615)

0.576 (0.556, 0.597)

0.891 (0.868, 0.911)

0.640 (0.604, 0.675)

0.791 (0.771, 0.809)

0.675 (0.651, 0.698)

0.496 (0.477, 0.515)

DIM-CG

0.668

0.605 (0.592, 0.618)

0.559 (0.540, 0.577)

0.885 (0.863, 0.906)

0.661 (0.625, 0.696)

0.785 (0.765, 0.804)

0.655 (0.630, 0.680)

0.470 (0.451, 0.488)

NOM-CG

0.656

0.602 (0.588, 0.616)

0.568 (0.547, 0.587)

0.885 (0.862, 0.905)

0.637 (0.603, 0.674)

0.739 (0.719, 0.760)

0.660 (0.635, 0.685)

0.457 (0.438, 0.476)

CONTRAfold2.0

0.656

0.651 (0.639, 0.664)

0.550 (0.532, 0.568)

0.869 (0.846, 0.891)

0.607 (0.569, 0.645)

0.746 (0.729, 0.763)

0.609 (0.587, 0.633)

0.509 (0.488, 0.527)

CentroidFold

0.643

0.642 (0.630, 0.654)

0.537 (0.517, 0.556)

0.860 (0.833, 0.885)

0.607 (0.568, 0.646)

0.705 (0.683, 0.724)

0.623 (0.600, 0.646)

0.492 (0.473, 0.512)

MaxExpect

0.625

0.577 (0.564, 0.589)

0.508 (0.488, 0.527)

0.858 (0.828, 0.883)

0.644 (0.611, 0.680)

0.695 (0.673, 0.715)

0.634 (0.608, 0.659)

0.435 (0.417, 0.452)

CONTRAfold1.1

0.601

0.590 (0.578, 0.602)

0.440 (0.421, 0.459)

0.841 (0.817, 0.866)

0.597 (0.565, 0.630)

0.690 (0.669, 0.712)

0.619 (0.594, 0.643)

0.392 (0.374, 0.410)

T99

0.597

0.546 (0.531, 0.560)

0.502 (0.481, 0.522)

0.860 (0.833, 0.885)

0.625 (0.594, 0.657)

0.583 (0.563, 0.604)

0.689 (0.666, 0.710)

0.389 (0.371, 0.406)

AveRNA

0.716

0.653 (0.641, 0.665)

0.618 (0.600, 0.638)

0.906 (0.884, 0.925)

0.683 (0.645, 0.719)

0.794 (0.776, 0.812)

0.732 (0.707, 0.753)

0.592 (0.575, 0.608)

AveRNA-I

 

0.676 (0.663, 0.687)

0.619 (0.602, 0.639)

0.901 (0.878, 0.922)

0.673 (0.640, 0.707)

0.808 (0.789, 0.825)

0.736 (0.715, 0.757)

0.590 (0.569, 0.608)

AveRNA-E

 

0.650 (0.637, 0.663)

0.617 (0.597, 0.637)

0.907 (0.885, 0.926)

0.683 (0.646, 0.718)

0.794 (0.774, 0.811)

0.710 (0.688, 0.733)

0.573 (0.555, 0.589)

  1. F-measure values for different algorithms for different classes in the S-STRAND2 dataset. AveRNA-I has been trained on 20% of the given class sampled uniformly at random, and the overall F-measure for the entire class is reported. AveRNA-E has been trained on 20% of S-STRAND2 excluding the given class, and the F-measure for the given class is reported.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us