Figure 2. Contributions to the solvation free energy of the G_B protein.

The protein was simulated in boxes of 3 different sizes (A). (B) Contribution to solvation $\mathrm{\Delta }G$ from charge introduction in water (left), vacuum (middle) and $\mathrm{\Delta }{G}_{water}-\mathrm{\Delta }{G}_{vacuum}^{aperiodic}$ (right). The latter term was calculated by considering a $\mathrm{\Delta }{G}_{water}$ value (i.e. free energy of switching on protein charges in water) for each of the box sizes (left panel) and $\mathrm{\Delta }{G}_{vacuum}$ (i.e. free energy of switching on protein charges in the gas phase) calculated in the infinite box without periodic boundaries (blue square in the middle panel). Blue markers denote those cases that are box size independent, while red symbols are for the box size-dependent contributions to $\mathrm{\Delta }G$. (C) Contribution to solvation $\mathrm{\Delta }G$ from switching on van der Waals interactions of the solute. In the panels (B) and (C) circles denote values obtained from individual free energy perturbation (FEP) calculations, triangle symbol marks those FEP runs where Hamiltonian replica exchange (HREX) was used, squares mark averages over the individual calculations. Upper x-axis marks the number of waters in the system.