Skip to main content

Table 2 Absolute and relative variations in O2PLS

From: Evaluation of O2PLS in Omics data integration

a

R X 2

R Y 2

R Xcorr 2

R Xcorr 2

R Xhat 2

R Yhat 2

R Xhat 2 / R Xcorr 2

R Yhat 2 / R Ycorr 2

1

57.97

50.81

46.31

1.37

26.74

0.80

57.74

58.55

2

67.94

53.40

60.80

4.24

29.52

1.45

48.55

34.25

3

74.08

54.79

68.99

7.35

26.70

2.00

38.69

27.23

4

78.06

55.62

72.94

9.63

29.23

2.40

40.07

24.87

5

80.93

56.69

76.51

11.30

29.81

3.32

38.97

29.43

 

1 E i , j 2 X i , j 2

1 F i , j 2 Y i , j 2

T W T i , j 2 X i , j 2

U C T i , j 2 Y i , j 2

U B U W T i , j 2 X i , j 2

T B T C T i , j 2 Y i , j 2

  
  1. The amount of variation per model statistic with respect to the total amount of variation, from an O2PLS fit using Metabolomics (X) and Transcriptomics (Y). The R 2 (definition given in last row) in percentages (with respect to the total variation in X and Y respectively) for each model statistic. The numbers of orthogonal components are n X =1,n Y =8. The number of joint components varies from 1–5. The first row was found best according to the proposed alternative cross-validation (as in Section “Methods”)