Skip to main content

Table 4 Validation analysis of five genes using five GEO datasets

From: Bayesian variable selection for parametric survival model with applications to cancer omics data

Gene symbol GSE14210 (N = 119) GSE15459 (N = 197) GSE29272 (N = 126) GSE51105 (N = 93) GSE62254 (N = 283) Heterogeneity-P (I2) Combined (random)
CTLA4 (2q33.2) a 0.77 (0.52~1.12)
1.7E−01
0.45 (0.25~0.82)
7.6E−03
0.59 (0.41~0.85)
4.1E−03
3.0E−01
16.9%
0.62 (0.48~0.82)
5.6E−04
PLCXD3 (5p13.1) 0.77 (0.51~1.15)
2.0E−01
1.60 (0.98~2.6)
5.8E−02
1.92 (1.32~2.79)
5.0E−04
4.0E−03
82.1%
1.33 (0.75~2.36)
3.3E−01
NACAD (7p13) 1.47 (0.97~2.24)
7.1E−02
1.47 (0.93~2.34)
9.8E−02
1.61 (1.04~ 2.49)
3.0E−02
1.76 (1.06~2.92)
2.7E−02
2.03 (1.41~2.92)
9.3E−05
7.7E−01
0.0%
1.68 (1.39~2.04)
1.2E−07
SERPINE1 (7q22.1) 1.46 (0.99~2.16)
5.6E−02
0.83 (0.56~1.24)
3.7E−01
0.73 (0.48~ 1.10)
1.3E−01
1.27 (0.76~2.12)
3.7E−01
1.57 (1.08~2.28)
1.6E−02
2.2E−02
65.0%
1.12 (0.82~1.53)
4.7E−01
GAMT (19p13.3) 1.17 (0.81~1.70)
4.0E−01
0.69 (0.46~1.03)
6.6E−02
1.81 (1.07~ 3.06)
2.6E−02
0.54 (0.32~0.92)
2.2E−02
1.75 (1.22~2.51)
2.2E−03
< 1.0E−03
81.8%
1.08 (0.69~1.68)
7.5E−01
  1. The contents of each cell represent estimated hazard ratio (HR), 95%CI of HR, and hypothesis testing P value of HR, and in each dataset, the samples are categorized into two groups using a best cutoff of expression level. Combined results are derived by meta-analysis with random effect model
  2. aExpression data of the gene is not available in the corresponding GEO database