We assessed the ability of three widely used and freely available programs for side chain repacking to simulate the structural effects of single point mutations. The programs tested seem to produce sufficiently reliable predictions only when mutations involve residues characterized by limited flexibility. The change in size and/or polarity of the mutant side chain can affect the performances of the different programs. The correlation between the quality of predictions, exposure to solvent, and B-factors of the corresponding residues in crystal is also investigated. This analysis may provide non-experts with insights into what types of modelling protocols work best for such a problem as well as providing information to developers for improving their algorithms.