Skip to main content

Table 1 Performance metrics of mRNA localization prediction using random forest (RF)

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 (1504)

73.24 ± 0.79

68.51 ± 1.36

70.87 ± 0.37

41.80 ± 0.71

71.53 ± 0.24

78.13 ± 0.20

77.43 ± 0.61

Cytosol (1798)

64.53 ± 0.61

72.11 ± 0.63

68.32 ± 0.16

36.75 ± 0.32

66.83 ± 0.56

75.63 ± 0.28

71.75 ± 0.49

EPR (850)

63.04 ± 1.72

73.68 ± 1.68

68.36 ± 0.99

36.95 ± 2.00

66.84 ± 1.27

75.54 ± 0.54

72.79 ± 0.77

Exosome (703)

63.20 ± 1.56

74.37 ± 1.13

68.79 ± 1.21

37.81 ± 2.41

66.74 ± 1.25

76.47 ± 0.62

77.57 ± 0.67

Mitochondrion (381)

98.53 ± 0.14

91.66 ± 7.95

96.46 ± 1.03

91.26 ± 5.76

95.59 ± 3.66

98.98 ± 0.20

99.27 ± 0.24

Nucleus (2754)

72.89 ± 0.92

73.99 ± 0.19

73.44 ± 0.54

46.88 ± 1.07

73.51 ± 0.60

80.28 ± 0.21

79.12 ± 0.27

Pseudopodium (180)

72.89 ± 1.20

69.00 ± 1.64

70.94 ± 0.97

41.93 ± 1.94

71.34 ± 1.02

76.73 ± 0.50

73.82 ± 1.62

Posterior (156)

98.19 ± 0.29

96.65 ± 0.84

97.42 ± 0.46

94.85 ± 0.90

97.19 ± 0.55

98.90 ± 0.83

98.29 ± 1.36

Ribosome (1532)

72.65 ± 1.01

70.89 ± 0.81

71.77 ± 0.24

43.55 ± 0.50

72.19 ± 0.51

78.40 ± 0.09

75.44 ± 0.29

  1. For each localization, a balanced dataset with equal number of positive and negative instances was used for prediction using RF. Performance metrics are computed following majority voting strategy, where 5 RF classifiers are constructed for each localization. Besides, accuracies in each classifier are measured following five-fold cross-validation
  2. EPR Endoplasmic reticulum