2.2 Selection of structural files of proteins bearing pathogenic and benign missense mutations.
To bring the analysis to the structural level we have queried VarMap [19], to associate each of our mutations to a protein structure file. Structures were available in the PDB only for a fraction of the mutations. Furthermore, as PISA database contains only 84% of all protein PDB entries, we had to take into account 1,253 and 1,580 protein structures to locate respectively 5,641 PMMs and 3,018 BMMs. The surface accessibility of replaced amino acids has been calculated with POPS software [20]. PyMOL [21] was used for molecular graphics and to generate mutant structures by using its Mutagenesis routine; we used the Optimize plugin for energy minimization of mutants [22].
2.3 Profiling of amino acid occurrences for pathogenic and benign missense mutations. We have implemented a series of Python scripts, downloadable from https://github.com [23] to collect all amino acid replacements that occur in BMM and PMM data sets.