Skip to main content

Table 3 Power estimates of MB-MDR to detect the correct interacting pair (SNP1, SNP2)

From: A robustness study of parametric and non-parametric tests in model-based multifactor dimensionality reduction for epistasis detection

  Trait status Power
g 2 Distributions Variances ST WT Rank_ST Rank_WT Log_ST Log_WT Rtn_ST Rtn_WT
  Normal Equal 0.400 0.046 0.367 0.001 0.377 0.039 0.378 0.041
  Normal Unequal 0.330 0.083 0.391 0.001 0.331 0.069 0.344 0.051
  Chi-square Equal 0.221 0.000 0.953 0.130 0.929 0.466 0.978 0.802
0.05 Chi-square Unequal 0.317 0.005 0.511 0.002 0.402 0.012 0.578 0.135
  t-distribution Equal 0.344 0.239 0.920 0.042 0.338 0.240 0.806 0.320
  t-distribution Unequal 0.383 0.116 0.615 0.002 0.380 0.122 0.543 0.132
  Normal Equal 0.950 0.634 0.952 0.087 0.959 0.626 0.958 0.650
  Normal Unequal 0.963 0.743 0.975 0.152 0.955 0.727 0.959 0.690
  Chi-square Equal 0.897 0.126 1.000 0.922 1.000 1.000 1.000 1.000
0.1 Chi-square Unequal 0.938 0.350 0.989 0.255 0.975 0.548 0.991 0.884
  t-distribution Equal 0.873 0.881 1.000 0.885 0.853 0.876 0.999 0.987
  t-distribution Unequal 0.921 0.801 0.995 0.409 0.921 0.806 0.989 0.834
  1. Legend Power is defined as the proportion of simulated samples of which the causal pair (SNP1, SNP2) is significant.
  2. ST=Student’s t-test, WT=Welch’s t-test, Rank_ST (Rank_WT)=Student’s t-test (Welch’s t-test) on trait ranks, Log_ST (Log_WT)=Student’s t-test (Welch’s t-test) on log transformed trait, Rtn_ST (Rtn_WT)= Student’s t-test (Welch’s t-test) on trait rank transformed to normal.