Skip to main content

Table 3 Per-segment performance for TMB (transmembrane beta strands). *

From: TMbed: transmembrane proteins predicted through language model embeddings

 

TMB (57/768)

Recall (%)

Precision (%)

Qok (%)

Qnum (%)

Qtop (%)

TMbed

95.0 ± 4.3

99.2 ± 0.7

80.5 ± 11.4

88.1 ± 6.9

98.1 ± 3.8

DeepTMHMM

85.9 ± 6.6

92.5 ± 4.7

46.1 ± 7.6

74.3 ± 13.0

97.2 ± 4.4

BOCTOPUS2

85.3 ± 9.2

96.6 ± 2.0

56.6 ± 18.9

71.2 ± 11.8

98.0 ± 2.0

BetAware-Deep

67.1 ± 6.5

62.2 ± 11.4

8.7 ± 5.3

60.9 ± 14.1

95.7 ± 5.4

PRED-TMBB2 (MSA)

85.4 ± 1.9

75.6 ± 4.8

18.4 ± 15.0

44.5 ± 26.7

95.9 ± 3.4

PROFtmb

78.2 ± 10.1

78.0 ± 6.9

20.2 ± 12.8

46.6 ± 11.7

97.2 ± 1.0

  1. *Segment performance for transmembrane beta strand (TMB) prediction based on 57 beta barrel TMPs (β-TMP) with a total of 768 TMBs. Recall, Precision, Qok, Qnum, and Qtop were averaged over the five independent cross-validation test sets; error margins given for the 95% confidence interval (1.96*standard error); bold: best values for each column; italics: differences statistically significant with over 95% confidence (only computed between best and 2nd best)