Skip to main content

Table 8 Predictive performances of the random forest models developed using relative abundance statistics along with the p-values corresponding to mean AUC values in 100 independent realizations (STD corresponds to standard deviation)

From: Predicting network modules of cell cycle regulators using relative protein abundance statistics

Positive class AUC (Mean ±STD) p-value AUC (Mean ±STD) p-value
    with randomized modules with randomized modules
START 0.8667 ±0.0004 <1.0E- 15 0.4996 ±0.0046 0.55
S/G2/M 0.8326 ±0.0005 <1.0E-15 0.5003 ±0.0038 0.46
EXIT 0.8366 ±0.0005 <1.0E-15 0.5008 ±0.0038 0.40
  1. Here, for each relative abundance, the network module of the cell cycle regulator in the “numerator” is used as the true class label of the relative abundance for model training and testing. The results were practically identical (less than 0.01 change in AUC values) when the regulator in the “denominator” was used as the true class label