Skip to main content

Table 2 Top 10 selected covariates for preclustered GO-groups according to 100 splits into training and test data

From: Survival models with preclustered gene groups as covariates

count

GO

effect

medoid

clustersize

annotation

85

GO:0043170

-1

209258_s_at

410

macromolecule metabolic process

81

GO:0007049

-1

210052_s_at

222

cell cycle

74

GO:0050896

+1

211908_x_at

102

response to stimulus

52

GO:0032501

+1

212195_at

310

multicellular organismal process

40

GO:0032501

+1

210935_s_at

362

multicellular organismal process

21

GO:0050794

+1

210417_s_at

312

regulation of cellular process

18

GO:0043170

-1

211693_at

434

macromolecule metabolic process

18

GO:0050896

+1

204118_at

230

response to stimulus

16

GO:0006952

+1

203535_at

27

defense response

15

GO:0042221

-1

219140_s_at

39

response to chemical stimulus

  1. Names are GenBank IDs for the medoid genes and GO IDs for GO groups [7]. The first column corresponds to the selected number for the covariate across 100 splits into training and test data for L 1 regression on the Mainz cohort study. The value of the effect indicates whether the covariate has an increasing(+1) or decreasing (- 1) effect on patients' risk to die.