Skip to main content

Table 1 Performance comparison among various methods.

From: Identifying network biomarkers based on protein-protein interactions and expression data

A      
  Data sets Method Predicting accuracy Redundancyscore Number of genes
Two-class case GSE10797 DEG+ellipsoidFN 93.94% (62/66) 0.2090 22
   PPIA+ellipsoidFN 96.97% (64/66) 0.1839 14
  GSE7904 DEG+ellipsoidFN 98.39% (61/62) 0.3698 18
   PPIA+ellipsoidFN 98.39% (61/62) 0.2968 19
Multiple class GSE18229 DEG+ellipsoidFN 80.00% (248/310) 0.1532 170
   PPIA+ellipsoidFN 79.03% (245/310) 0.1652 135
  GSE10797 DEG+ellipsoidFN 78.79% (52/66) 0.1663 161
   PPIA+ellipsoidFN 83.33% (55/66) 0.2150 56
B      
  Data sets Method Redundancy score Number of common genes
Two-class case GSE10797 t-test 0.2715 13
   PPIA 0.2144   
  GSE7904 t-test 0.3077 27
   PPIA 0.2623   
Multiple class GSE18229 F-test 0.3031 15
   PPIA 0.1693   
  GSE10797 F-test 0.3280 11
   PPIA 0.2209   
  1. (A) Comparing PPIA + ellipsoidFN method with DEG + ellipsoidFN in predicting accuracy, redundancy score, and the number of genes identified. (B) Comparisons on PPIA + ellipsoidFN and t-test for two-class case and PPIA + ellipsoidFN versus F-test for multiple-class case based on 50 top PPIs and proteins.