Detecting noncrystallographic symmetry in Icosahedral Viruses using deep learning approach

Assembly of capsid virus is a crucial step in virus life cycle. Without this step, virus will not replicate itself to hijack other cells and its life cycle would end. Many researchers studied virus structural shape and its dynamics to understand the behavior of the virus. So, this paper focuses on the structural shape of Icosahedral viruses and prediction of symmetries in their capsids. A small virus capsid contains identical asymmetric units that are packed in regular manner. Every icosahedral virus has two types of symmetry, regular symmetry and non-crystallographic symmetry. So, one asymmetric unit and some rotation matrices are needed to form the whole capsid. These rotation matrices define the location of adjacent asymmetry unit. In this paper, deep learning approach is followed to create a layered model that predicts non-crystallographic symmetry in virus capsid. Through visualization technique, the results were promising; the accuracy was 89% for assembling the capsid in icosahedral viruses using dataset taken from the Protein Data Bank (PDB).