Leveraging structural relationships as a novel mode of viral classification

Abstract

100 years have passed since the independent discovery of the humble bacteriophage (phage) by Frederick Twort and Felix d’Herelle in 1915 and 1917 respectively, and since then, it has become commonly accepted that phages represent the most abundant biological entities on Earth. Despite this fact, viral taxonomy lies in extremely treacherous waters, ever changing to accommodate the next series of phylogenetic mysteries. The utilisation of genes such as the terminase large sub-unit can in some cases provide a robust taxonomic marker, but this is often found to fail at higher taxonomic levels. In addition, the rapid evolutionary dynamics and highly modular nature of phages provide yet more phylogenetic roadblocks, necessitating additional and multifaceted approaches as a means of resolution. Here, we describe a novel approach towards the taxonomic classification of phage systems. Tools for accurately predicting the three dimensional structure of proteins are improving at an unprecedented rate due to the fact that the number of protein sequences far exceeds the number of experimentally determined structures. Our approach leverages these methods through a pipeline which compares models of phage marker genes in order to permit the inference of phylogenetic relationships based on cross model superimposition. We hope this method will supplement other approaches in providing a more holistic approach to viral classification.