Table 4.
Ten Most Significant Predictors for Predicting Survival Outcome in Sweden Breast Cancer Dataset (Miller et al., 2005) Using Pathway-Based Supervised PCA Model
| Gene set | Description | Set size | No. of selected genes | Gene symbols |
|---|---|---|---|---|
| GO:0019233 | Sensory perception of pain | 5 | 4 | NPY1R, NGF, NDN, PENK |
| GO:0007193 | G-protein signaling, adenylate cyclase inhibiting pathway | 10 | 6 | RGS1, EDG1, NPY1R, 0PRD1, CORT, GPR44 |
| GO:0030520 | Estrogen receptor signaling pathway | 9 | 8 | TAF7, RBM14, ESR1, NC0A6, ARID1A, ESR2, RBM9, DDX54 |
| GO:0050729 | Positive regulation of inflammatory response | 5 | 3 | STAT5A, FABP4, STAT5B |
| GO:0007631 | Feeding behavior | 20 | 12 | NPY1R, BDNF, AGRP, HTR2C, HCRTR1, HCRT CCKBR, CCKAR, GALR2 PYY, FYN, MCHR1 |
| GO:0001816 | Cytokine production | 8 | 7 | CD4, FABP4, PTAFR, CD226, MAF, NFATC2IP, N0D2 |
| GO:0007026 | Negative regulation of microtubule depolymerization | 10 | 4 | MAP4, MAPT, CLASP2, MID1IP1 |
| GO:0007626 | Locomotory behavior | 17 | 13 | SEPP1, HEXA, HEXB, GNA01, NPY1R, C1QL1, N0VA1, PGDS, GRM1, ADRA1B, NRG1, DRD1, CELSR1 |
| GO:0007187 | G-protein signaling, coupled to cyclic nucleotide second messenger | 37 | 29 | VIPR1, NPY1R, MC1R, PTHR1, NPY, CNR2, HTR6, HTR1E, HTR, 0PRD1, HTR7, SSTR1, MTNR1B, MC5R, HTR1B, CALCRL, 0PRM1, CNR1, SSTR4, DRD1, CCL2, SSTR2, HRH3, HRH2, EDG7, MTNR1A, MC3R, HTR1F, XCR1 |
| GO:0032012 | Regulation of ARF protein signal transduction | 13 | 6 | PSD4, PSD3, PSD, IQSEC3, ARFGEF2, PSCD4 |
“Set size” refers to the number of genes in the GO category. In addition to identifying gene categories associated with outcome, the pathway-based Supervised PCA model also carries out additional within-category selection to select important genes within each gene set. “No. of selected genes” refers to the number of selected genes in each gene set that were used to estimate principal component score or supergene. Gene symbols for the selected genes are listed in the last column.