2.4.4 | Prior-checking of simulations’ fit to the ACB and ASW data
We plotted each prior summary statistics distributions and visually verified that the observed summary statistics for the ACB and ASW respectively fell within the simulated distributions (Supplementary Figure S7 ). Then, we explored the first four PCA axes computed with the princomp function in R , based on the 24 summary statistics and all 90,000 simulations, and visually checked that observed summary statistics were within the cloud of simulated statistics (Supplementary Figure S8 ). Finally, we performed a goodness-of-fit approach using the gfit function from the abc package in R , with 1,000 replicates and tolerance level 0.01 (Supplementary Figure S9 ).