Skip to main content

Table 13 Comparison of structures with Proteinnet DSSP3 dataset

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

Methods

TEST2018

SPOT-2016

SPOT-2016-HQ

SPOT-2018

SPOT-2018-HQ

AttSec(this work)

86.6

81.6

81.9

81.5

81.3

Transformer only

86.0

81.4

81.6

81.3

80.9

LM’s attention

85.2

80.1

80.4

80.0

80.0