Understanding the protein capsid structure of a virus through rigidity and flexibility analyses through computational methods, such as those provided by KINARI-Web, can deliver crucial insight into the function of viruses. However, due to the immensely large size of many viral capsids of interest, currently, there is no time and space-efficient method to analyze full capsid assemblies. Thus, it is of interest to use geometric principles of viral capsids, which are made of repetitive asymmetric units, to eliminate unnecessary computation, and pre-process virus files for visual examination and analysis.
For my thesis, I have created an assembly scaffold that takes a viral asymmetric unit from the Protein Data Bank (PDB) and applies algorithms rooted in geometric principles to build a connectivity graph for a viral assembly in two steps. First, using convex hulls and breadth first search algorithms, I created a graph that captures how asymmetric units are connected once assembled. Then, by checking for where chains of asymmetric units overlap, I annotated this connectivity graph to complete the scaffold. By applying the pipeline to several viral examples, I demonstrated that these tools provide novel insights into this kind of computational data. Through building a scaffold, future extensions of my pipeline will reconstruct connections between capsomeres of asymmetric units and ultimately, prepare large viral data for rigidity and flexibility analysis.
A poster derived from honors research with Ileana Streinu, Charles N. Clark Professor of Computer Science.