The root mean square error between predicted and measured ΔΔG values in the validation sets (σc) is given as a function of ΔΔGP (continuous line) and ΔΔGM (dashed line). The dataset of mutations was divided in non-overlapping subsets corresponding to intervals of ΔΔGP (or ΔΔGM) of 0.25 kcal/mol width, and σc was computed on each subset separately using eq. 4. Subsets containing less than 10 mutations were dismissed.
Dehouck et al. BMC Bioinformatics 2011 12:151 doi:10.1186/1471-2105-12-151