Skip to main content

Table 3 Recall of individual secondary structures is compared on CB513 for Q8 prediction

From: Prediction of 8-state protein secondary structures by a novel deep learning architecture

Q8 Label

CRRNN

NCCNN

MUFOLD-SS

DCRNN2a

DCRNN

DeepCNF

H

0.926

0.932

0.920

0.920

0.933

0.904

B

0.081

0.041

0.071

0.003

0.026

0.026

E

0.831

0.821

0.815

0.841

0.828

0.833

G

0.371

0.285

0.364

0.359

0.252

0.26

I

0

0

0

0

0

0

T

0.555

0.524

0.549

0.539

0.522

0.528

S

0.332

0.24

0.290

0.258

0.249

0.255

L

0.658

0.69

0.662

0.658

0.652

0.657

  1. aData is generated by our experiment
  2. Boldface numbers indicate best performance