The extensive genomic variation within cancer cells makes it hard to distinguish driver and passenger mutations, but can have significant implications for disease progression and treatment. This is further complicated by the broad range of potential molecular consequences of coding mutations, including how they affect a protein's structure, function and interactions. Understanding these effects can provide invaluable insight into the behavior and drug resistance of cancer cells. While there are many tools available to analyze the effects of coding mutations on proteins, these are all limited by either not providing any mechanistic insight, or by focusing on specific molecular consequences. In order to fully interpret the effect of a mutation, one must consider the full range of potential molecular consequences.
Here we present a novel integrated tool for the analysis of coding missense mutations, which facilitates the analysis of the effects of mutations on all aspects of the protein structures and functionalities using available structural information. The server accepts a protein structure file to analyse either a given list of mutations, or to perform saturation alanine mutagenesis. We then leverage the tools SDM2, mCSM-Stability and DUET to assess the impact of each mutation on protein folding and stability; mCSM-PPI, mCSM-NA2 and mCSM-lig to analyse the effect of each mutation on protein interactions; which are combined with structural analyses of conservation, electrostatics, flexibility calculated using Encomm and molecular interactions calculated using Arpeggio. The results are presented through an interface that provides a machine learned interpretation of the results based upon the predictions normalised to the protein of interest. The raw data along with pymol session files to enable ready visualisation of each result are made available for download. This has allowed us to distinguish between gain of function and loss of function p53 mutations, and to predict clinical outcomes for patients with mutations in VHL and the succinate dehydrogenase complex.
This presents the first fully integrated structural based mutation analysis platform.