From: Discovery of stroke-related blood biomarkers from gene expression network models
Method | Number of biomarkers | Predictive Accuracy (results with 5-fold cross validation) |
---|---|---|
Differential Expression analysis | 557 | < 71% |
Support Vector Machines (default parameters) using the10 most differentially expressed genes as input | 10 | 75.1% |
InSyBio predictive analytics approach using differentially expressed gene set | 25 | 81.21% |
Gene expression signature from O’Connell et al. 2017 | 10 (ANTXR2, STK3, PDK4, CD163, MAL, GRAP, ID3, CTSZ, KIF1B, PLXDC2) | 82.07% |
InSyBio predictive analytics using network significant gene set | 6 (ID3, MBTPS1, NOG, SFXN2, BMX, SLC22A1) | 89.57% |