Skip to main content

Table 4 Comparative analysis of the developed approach with the mRNALoc

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

 

mRNALoc

Proposed

Sn

Sp

Ac

F1-score

MCC

Sn

Sp

Ac

F1-score

MCC

Test set-I

Cytoplasm (86)

39.5

75.6

57.5

44.1

16.2

76.7

59.0

67.9

63.1

36.1

Endoplasmic reticulum (31)

51.7

88.8

70.2

46.6

42.1

44.8

91.8

68.3

45.7

40.2

Mitochondrion (25)

48.0

93.8

70.9

48.6

47.5

56.0

73.5

64.8

30.4

22.3

Nucleus (83)

50.0

62.2

56.1

46.1

12.3

39.0

82.3

60.7

46.6

23.8

Test set-II

Cytoplasm (464)

52.6

64.5

58.6

52.6

17.2

63.4

65.3

64.3

60.5

28.6

Endoplasmic reticulum (103)

43.7

78.4

61.0

25.0

17.3

72.8

93.1

83.0

61.2

62.4

Mitochondrion (8)

75.0

97.8

86.4

65.2

47.7

100

82.7

91.3

31.6

21.4

Nucleus (508)

55.5

78.7

67.1

61.8

34.4

55.1

100

77.6

71.1

58.6

  1. For the Test set-I, the developed method achieved higher accuracy than the mRNALoc for cytoplasm and nucleus localizations. For the Test set-II, the developed method performed better than mRNALoc in all the four localizations
  2. Bold font denotes the higher performance of the proposed approach as compared to the mRNALoc
  3. Sn sensitivity, Sp specificity, Ac accuracy, MCC Matthew’s correlation coefficient