Impact of error and replicas overlap on \(\Delta G\) calculations. (A) The dataset is sorted based on the value of the largest error of the two \(\Delta G\) calculations which determines each \(\Delta\Delta G_R\) estimate (see Eq. 1 in the main text). In red we report this sorted error, in red the corresponding running average for \(\Delta\Delta G_R\) and in cyan the \(\Delta\Delta G_R\) for each mutation, labelling the 15 ones with the largest \(\Delta\Delta G_R\). (B) Average error on \(\Delta\Delta G_R\)calculations subdividing the dataset in the different categories analyzed in Fig. 5 of the main text. (C) Average exchange rejection (AER) subdividing the dataset in the different categories analyzed in Fig. 5 of the main text.