Application of Methods from Information Theory in Protein-Interaction Analysis

Abstract

The interaction of proteins with other biomolecules plays a central role in various aspects of the structural and functional organization of the cell. Their elucidation is crucial to understand processes such as metabolic control, signal transduction, and gene regulation. However, an experimental structural characterization of all of them is impractical, and only a small fraction of the potential complexes will be amenable to direct experimental analysis. Docking represents a versatile and powerful method to predict the geometry of protein–protein complexes. However, despite significant methodical advances, the identification of good docking solutions among a large number of false solutions still remains a difficult task. The present work allowed to adapt the formalism of mutual information (MI) from information theory to protein docking. In this context, we have developed a method, which finds a lower bound for the MI between a binary and an arbitrary finite random variable with joint distributions that have a variational distance not greater than a known value to a known joint distribution. This lower bound can be applied to MI estimation with confidence intervals. Different from previous results, these confidence intervals do not need any assumptions on the distribution or the sample size. An MI-based optimization protocol in conjunction with a clustering procedure was used to define reduced amino acids alphabets describing the interface properties of protein complexes. The reduced alphabets were subsequently converted into a scoring function for the evaluation of docking solutions, which is available for public use via a web service. The approach outlined above has recently been extended to the analysis of protein–DNA complexes by taking also into account geometrical parameters of the DNA.