Table 2 Simulation results for experiment-wide 1%, 5% and 10% significance thresholds compared to thresholds based on Bonferroni corrections using the calculated number of independent comparisons