Predicting protein thermal stability changes upon point mutations using statistical potentials: Introducing HoTMuSiC

• DOI: 10.1038/srep23257
• Publication Date: 2016-03-18T10:13:53Z
• Authors: Fabrizio Pucci; Raphaël Bourgeas; Marianne Rooman
• Keywords: Root mean square; Thermal Stability

Abstract The accurate prediction of the impact an amino acid substitution on thermal stability a protein is central issue in science, and key relevance for rational optimization various bioprocesses that use enzymes unusual conditions. Here we present one first computational tools to predict change melting temperature Δ T m upon point mutations, given structure and, when available, wild-type protein. ingredients our model are standard temperature-dependent statistical potentials, which combined with help artificial neural network. was chosen basis detailed thermodynamic analysis system. parameters were identified set more than 1,600 mutations experimentally measured . performance method tested using strict 5-fold cross-validation procedure, found be significantly superior competing methods. We obtained root mean square deviation between predicted experimental values 4.2 °C reduces 2.9 ten percent outliers removed. A webserver-based tool freely available non-commercial at soft.dezyme.com