Size Matters: Finding the Perfect Probability Cutoff Value for Significance in Host Parasite Protein-Protein Interaction Networks

Abstract

The complex mechanisms underlying all biological function are governed by our genes and the proteins they code for, interacting with one another forming the basis of all life processes. The study of these protein-protein interactions is a well defined field, advances in which can not only help us understand the enigmatic foundations of life itself, but how perturbations in these interactions, pathogenic or mutative, can cause these systems to fail. With this knowledge, it is often the case that it can be better understood how to fight infections and genetic mutations in affected peoples.

However, these proteomic systems are often far too complex to be comprehensively measured, and it is therefore often difficult to meaningfully discern which interactions are significant and worth future study. This research works to outline and demonstrate a method for using basic graphical analyses to provide a specific cutoff point and delineate a specific subset of protein-protein interactions relevant to a system. This study, for the HCMV HSV-1 and KSHV obligate intracellular parasites, finds that pruning interactions in a network with a lower than 0.35 probability of occurring reduces the network to a manageable size while also maintaining all or most of its emergent structure and function for research. The research also explores several properties of the networks analyzed that can be gleaned from such graphical analysis, as well as explores aspects of the networks that differ from expectation and from one another.