Skip to main content

Table 2 Average prediction accuracy of models trained with Proteinnet DSSP8 dataset

From: AttSec: protein secondary structure prediction by capturing local patterns from attention map

Model

TEST2018

SPOT-2016

SPOT-2016-HQ

SPOT-2018

SPOT-2018-HQ

PSIPRED-Single

–

–

–

–

–

SPIDER3-Single

59.8

58.9

59.9

57.4

58.0

ProteinUnet

60.3

–

–

-

–

SPOT-1D-Single

62.2

61.4

61.6

60.1

60.0

SPOT-1D (profile)

75.4

69.3

71.7

67.4

70.5

AttSec(ours)

77.4

70.4

73.0

71.2

72.3

SPOT-1D-Single (less than 1024)

65.5

64.1

65.8

64.6

64.8

SPOT-1D-LM (less than 1024)

77.5

70.0

72.6

70.9

72.3

AttSec(ours) (less than 1024)

77.4

70.5

73.0

71.4

72.4

  1. The best results were written in boldface