Comparison of the procedure based on multigene normalization followed by single-model MCMC analysis (“classic” model) to full Bayesian models.

<p>(a) Point-estimates and credible intervals for the effects of stress in coral dataset inferred by the informed model (allowing control genes <i>rpl11</i> and <i>nd5</i> to change 1.2-fold on average) and “classic” model based on normalization using the same two control genes. Points are posterior means, the whiskers denote 95% credible intervals. (b) Comparison of the power of the “classic” model (transformed p-values along the horizontal axis) to the power of full Bayesian models (see legend). The colored lines are linear regressions to illustrate trends (no statistical implications intended), the black dotted line is 1∶1 correspondence. “Classic” model generates the the credible intervals that are as narrow as under the fixed model, but relies on more realistic assumptions.