Skip to main content

Table 1 Prediction performance for different tools before filtering

From: Predicting the functional repertoire of an organism from unassembled RNA–seq data

Tool

Sensitivity

AUC (in %)

 

(in %)

count

sum–of–

means–

scaled mean–

   

scores

core

score

BLASTX

97.21

92.02

93.09

94.96

94.95

RAPSearch

97.04

96.13

96.26

95.38

96.78

PAUDA

96.36

96.89

96.94

94.66

97.07

UProC

97.10

92.20

95.62

95.68

96.03

 

Maximum F1–Score (in %)

 

count

sum–of–

mean–

scaled mean–

   

scores

score

score

BLASTX

74.19

76.77

84.60

82.80

RAPSearch

87.03

87.72

86.13

89.50

PAUDA

88.30

88.63

83.09

89.28

UProC

75.01

87.67

88.60

89.93

  1. The area under the curve (AUC) was calculated on sorted functions. The maximum F1–Score corresponds to the best possible separation between false and true predictions. Quality scores are averaged over all samples. The maximum AUC and F1–Score per tool are marked in bold text.
Back to article page

BMC Genomics

ISSN: 1471-2164

Contact us