Skip to main content

Table 2 Similar to Table 1, but this comparison is based on the second set of simulation studies (with subject effect)

From: A full Bayesian partition model for identifying hypo- and hyper-methylated loci from single nucleotide resolution sequencing data

   Logistic DSS z-test Bayesian
   regression    
γ=0.01 FDR 0.0027 0 0.0034 0.0069
  mdFDR 0.0029 0 0.0036 0.008
  TPR 0.1953 0.0171 0.183 0.2494
  TPR_hypo 0.1948 0.0167 0.1918 0.2545
  TPR_hyper 0.1958 0.0176 0.171 0.2418
γ=0.05 FDR 0.0097 0 0.0113 0.014
  mdFDR 0.0097 0 0.0113 0.0141
  TPR 0.3554 0.0364 0.3235 0.4038
  TPR_hypo 0.3604 0.0375 0.3392 0.419
  TPR_hyper 0.3494 0.035 0.3044 0.3853
γ=0.1 FDR 0.0119 0 0.0142 0.0156
  mdFDR 0.0119 0 0.0142 0.0156
  TPR 0.5053 0.0668 0.4588 0.5745
  TPR_hypo 0.5021 0.0692 0.4637 0.5836
  TPR_hyper 0.509 0.0641 0.453 0.5637
γ=0.15 FDR 0.012 0 0.0147 0.0156
  mdFDR 0.012 0 0.0147 0.0156
  TPR 0.6385 0.1074 0.5939 0.7158
  TPR_hypo 0.6309 0.112 0.5905 0.7205
  TPR_hyper 0.6477 0.1016 0.5981 0.7101
γ=0.2 FDR 0.0116 0.0001 0.0146 0.0153
  mdFDR 0.0116 0.0001 0.0146 0.0153
  TPR 0.7546 0.1576 0.7203 0.8254
  TPR_hypo 0.7459 0.1612 0.7129 0.8287
  TPR_hyper 0.7659 0.1529 0.7298 0.821