Skip to main content

Table 3 Accuracy, sensitivity, PPV and F-Score in % for Ion Torrent dataset and fold change threshold f of 1, 1.5, and 2.

From: Bootstrap-based differential gene expression analysis for RNA-Seq data with and without replicates

Fold Change Method Accuracy % Sensitivity % PPV % F-Score %
  FisherTotal 71.68% 72.76% 90.56% 80.69%
  FisherHousekeeping 67.15% 66.87% 94.74% 78.40%
  FisherERCC 71.39% 72.45% 88.97% 79.86%
1 GFOLD 75.77% 77.55% 90.43% 83.50%
  IsoDE-Match 81.75% 86.38% 82.18% 84.05%
  IsoDE-All 81.46% 86.07% 82.13% 84.05%
  FisherTotal 74.16% 78.39% 85.06% 81.59%
  FisherHousekeeping 76.06% 73.23% 92.96% 81.93%
  FisherERCC 74.31% 78.59% 85.45% 81.87%
1.5 GFOLD 75.47% 77.63% 87.88% 82.44%
  IsoDE-Match 77.66% 83.94% 84.75% 84.34%
  IsoDE-All 77.81% 84.13% 84.45% 84.29%
  FisherTotal 79.71% 83.02% 84.00% 83.51%
  FisherHousekeeping 81.75% 80.70% 88.75% 84.53%
  FisherERCC 79.42% 82.56% 84.12% 83.33%
2 GFOLD 80.58% 76.74% 90.66% 83.12%
  IsoDE-Match 81.75% 85.81% 84.63% 85.22%
  IsoDE-All 81.61% 86.28% 84.13% 85.19%
  1. The number of bootstrap samples is M = 200 for IsoDE-Match and M = 20 for IsoDE-All, and bootstrap support was determined using the binomial model with significance level α = 0.05.