Figure 5From: BGX: a Bioconductor package for the Bayesian integrated analysis of Affymetrix GeneChipsAdapting the size of MCMC proposal steps. When the adaptive MCMC algorithm is used, the variance of the Metropolis-Hastings proposal step is adapted independently for each S component. The plot shows a histogram of the optimal log variance for S proposals and the fixed step size used in the non-adaptive version overlaid in black, highlighting that a wide range of proposal variances are needed.Back to article page