Skip to main content

Table 3 Prediction accuracies with Independent test set-II

From: mLoc-mRNA: predicting multiple sub-cellular localization of mRNAs using random forest algorithm coupled with feature selection via elastic net

Localization

Performance metrics

Sensitivity

Specificity

Accuracy

MCC

F1-score

aucROC

aucPR

Cytoplasm (490)

91.84

66.57

69.34

37.26

39.66

91.77

73.72

Cytosol (1037)

89.78

71.73

75.91

52.49

63.39

92.21

83.36

EPR (485)

91.55

72.88

74.91

42.25

44.21

92.00

69.50

Exosome (185)

88.65

77.65

78.10

30.44

25.12

92.29

64.38

Mitochondrion (14)

92.86

95.29

95.28

22.62

10.99

98.97

49.99

Nucleus (1266)

90.76

76.47

80.51

61.25

72.54

93.00

86.41

Pseudopodium (79)

93.67

69.02

69.45

17.68

9.79

92.93

59.72

Posterior (121)

100.00

92.97

93.16

51.33

44.20

98.11

53.66

Ribosome (789)

91.48

73.73

76.89

51.45

58.28

93.05

80.26

  1. The sensitivities are found to be much higher than that of Independent test set-I. Because, this dataset shares > 80% sequence similarity with the training set. However, overall accuracies are found at par with the training dataset. It can also be seen that the aucROC values are much higher, may be due to higher degree of sequence similarity with the training set
  2. EPR Endoplasmic reticulum
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us