Table 2

Superpc continuous prediction results from breast cancer data analysis

Principal component


1st

2nd

3rd


NEJM

HR

5.21E+00

2.28E+00

1.98E+00

95% HR CI

3.18 to 8.56

1.19 to 4.34

1.02 to 3.83

P-value

6.68E-11

1.25E-02

4.26E-02

GSE4922

HR

7.35E+00

2.05E-01

2.66E-02

95% HR CI

2.11 to 25.6

0.0091 to 4.62

6.5e-4 to 1.08

P-value

1.73E-03

3.19E-01

5.51E-02

GSE3143

HR

7.27E+02

6.10E-03

1.70E+00

95% HR CI

8.1 to 6.5e+4

1.8e-4 to 0.21

1.1 to 2.6

P-value

4.12E-03

4.61E-03

1.19E-02

GSE18229

HR

5.34E+00

7.20E+00

9.31E-01

95% HR CI

2.63 to 10.9

1.73 to 30.0

0.30 to 2.90

P-value

3.63E-06

6.66E-03

9.01E-01

GSE1456

HR

1.50E+02

1.45E+02

6.03E+00

95% HR CI

17 to 1287

1.66 to 1.26e+4

0.034 to 1057

P-value

5.00E-06

2.91E-02

4.96E-01


The results were generated by using the Nejm data set as the training data set and four independent data sets as validation data with a threshold value of 1.20 and 8 selected MCL modules. P-values less than 0.05 are in bold. CI, confidence interval; HR, hazard ratio.

Wu and Stein Genome Biology 2012 13:R112   doi:10.1186/gb-2012-13-12-r112

Open Data