Table 4 Genes selected by different approaches
From: Dynamic incorporation of prior knowledge from multiple domains in biomarker discovery
Method | Parameters | Number of Selected Genes | Selected Genes | |
|---|---|---|---|---|
Lasso | λ = 0.03 | 15 | ATP5J, AURKA, GNPTAB, GPR137B, HSD17B4, IFNGR2, IGFBP5, MED30, MFF, SDC2, SMARCC1, TAF2, TUBB, UBE2J2, ZHX1 | |
RRF | γ = 0.7 | 13 | ARID4A, CASP3, CAV1, CCND1, CCNH, CDC25C, CDK10, FGF8, IGFBP5, MEN1, MMP3, PDGFB, SEMA3F | |
Know-GRRF (t = 10, c = 0.5) | Cancer gene prior | δ = 1 | 169 | Omitted due to space limits. |
Consv. prior | δ = 0.5 | 9 | BCL2L1, BMP4, COL1A1, E2F1, FAS, MEN1, PLAT, RAD23A, TSG101 | |
VI prior | δ = 0.2 | 6 | CSF2, DDX6, JUND, MMP3, NOTCH4, PURA | |
All priors | δ = 0.5 βdriver = 0.3 βconsv = 0.6 βvi = 0.1 | 7 | BMP4, CCNA2, FAS, MEN1, PTPRF, RAD23A, TSG101 | |