Amino acid characteristics in protein native state structures

Proteins are chain molecules made up of twenty naturally occuring amino acid types. The set of amino acid side chains span a range of geometrical shapes, physical sizes, as well as chemical properties. We analyze experimental data on amino acid identity and side chain orientations with respect to the protein backbone in more than 4000 high-resolution protein native state structures. We consider the magnitude and direction of protrusion along with a simplified chemical attribute captured by a hydrophobic scale. The geometrical, physical, and chemical properties of the side chains are simultaneously studied using principal component analysis to infer the consensus propensity to harmoniously substitute one amino acid with another in protein sequences, while preserving the native state structure. This propensity also indicates the destabilization potential of a point mutation to the native state fold. We benchmark our results against mutational studies reported in the literature.