Protein-protein interactions: Basics, characteristics, and predictions

Abstract

Most of the cellular processes involve protein-protein interactions (PPIs). It therefore necessitates obtaining the detailed information about the amino acid residues involved in PPIs. Available are the different PPI determining experimental techniques. These experimental methods, though very accurate, are time consuming, labor intensive, and very expensive. To solve the aforementioned problems, different labs developed different bioinformatic protocols to build different number of bioinformatic software tools to predict PPIs. The bioinformatic algorithms are used for prediction of three-dimensional structures of proteins as well as protein complexes. Nowadays, different machine learning algorithms are employed for the purpose of prediction of PPIs. The computational structure prediction methods involve homology modeling, threading, and ab initio modeling. These methods have nearly 75%-80% overall accuracies. The other most widely used method is molecular docking which is used to generate the three-dimensional conformations of protein complexes. The docking methods can broadly be categorized as rigid body docking and flexible docking. In this chapter, the different aspects of computational modeling and docking strategies will be covered. The basic terminologies will be revisited. The chapter will aim at providing a firsthand guide on protein interaction prediction methods.