Computational Prediction Approaches for Predicting Mutation Impact on Protein-Protein Interactions

Protein-protein interactions (PPIs) involves in many significant mechanisms for human. The mutation impacts on PPIs can lead to the differences in conformational stability of proteins, as well as the PPIs kinetics and thermodynamics. Revealing these impacts is essential for understanding the underlying mechanism and designing new therapies. Therefore, there is a need to develop a reliable predictor for free energy changes of protein-protein binding affinity upon mutants. There have been many methods built on this purpose in the past two decades. So, we aim to conclude several aspects of these computational predictors for mutation impacts on PPIs, including several kinds of features, databases and measures used by predictors. A comparison is also conducted but not so accurate as the performances on measures are not on the same dataset. So, we got a conclusion that a benchmark database is needed for general validation and comparison of approaches related to mutant impact on PPIs. Besides, we categorized previous computational predictors into three groups, named energy-based approaches, structure-based methods and sequence-based methods. After comparison, sequence-based methods are a bit better than structure-based methods on the measure of PCC, especially the method SAAMBE-SEQ and MUPIPR. This comparison also illustrates a trend to predicting the impact of mutations on more PPIs with only sequence information.