Software Pipeline for Predicting the Impact of Mutations on the Stability of Protein Spatial Structures Using Free Energy Change Estimation Methods and Artificial Intelligence

In this work, a computational pipeline was developed to predict the impact of mutations on the stability of protein structure. The pipeline employs a combined approach, utilizing state-of-the- art artificial intelligence methods for protein structure prediction and classical algorithms for free energy change estimation. The pipeline includes protein structure prediction using the ESM3 model, followed by calculation of free energy changes in mutant forms using pyRosetta. This approach allows overcoming the limitations of existing methods by combining the advantages of deep learning and the interpretability of energy calculations. The developed tool can find applications in structural bioinformatics, biotechnology, and medicine, especially given the limited number of experimentally determined protein structures.