Skip to main content

Table 4 Number of combinations that are also statistically significant using the validation data

From: Identifying statistically significant combinatorial markers for survival analysis

(a) Validated BRCA results

Expression

GSE2034

GSE25066

GSE3494-GPL96

GSE3494-GPL97

High

54/239 (22.59%)

37/172 (21.51%)

74/286 (25.87%)

23/106 (21.70%)

Low

466/2092 (22.28%)

404/2073 (19.49%)

421/2079 (20.25%)

105/670 (15.67%)

Total

509/2092 (24.33%)

428/2073 (20.65%)

485/2079 (23.33%)

123/670 (18.36%)

(b) Validated OV results

Expression

GSE13876

GSE49997

 

High

15/300 (5.00%)

15/108 (13.89%)

 

Low

195/1526 (12.78%)

140/1444 (9.70%)

 

Total

209/1526 (13.70%)

155/1444 (10.73%)

 
  1. Percentage values indicate the portion of statistically significant combinations from all combinations that can be matched in the data set. ‘High’ results use the binarization similar to our experiment settings where z-scores greater than 2 were set to 1, otherwise, 0. ‘Low’ results consider the case when the genes of interest are all lowly expressed and entries with z-scores less than − 0.5 are set to 1, otherwise, 0. All combinations were tested using these two binarizations, hence, the respective lists of statistically significant combinations found in the validation data may overlap. The ‘Total’ indicates the total number of unique combinations (high or low)that can be matched in the data and the portion of which are statistically significant