Speaker
Description
Visualizing flu substitutions may help identify the features that enable certain clades to succeed as others die out. For example, flu viral escape from prior antibody responses is most likely if mutations are on the virus surface proteins (hemagglutinin, HA, and neuraminidase, NA) and decrease antibody binding. We present a tool, Flu Strain Compare, which maps flu mutations onto 3D HA and NA proteins for visualization and to facilitate interpretation of mutational effects. As well as point mutations, changes in potential N-linked glycosylation sites (PNGS) between strains are displayed. PNGS can mask the protein surface from recognition by antibodies, thus it is expected that new transmissible strains will acquire PNGS. The user can elect whether to focus solely on non-conservative changes (amino acid changes that are likely to affect biochemical properties), changes at published epitopes or solvent-accessible sites; or the user can list sites to mask or highlight. Multiple strains can be compared, with coloring indicating diversity across the set, and multiple diversity indices have been implemented (e.g., Shannon entropy and Gini-Simpson diversity). The tool has been tested with the H1, H3, and H5 hemagglutinins and N2 and N1 neuraminidases.
The program takes as input amino acid sequences and a configuration file edited by the user, it runs rapidly on command line, using PyMOL and python. We illustrate tool functionality and utility by comparing H3N2 vaccine strains in H3 years from 2010 to 2019 to the clades that became dominant in those matched years. We determine whether vaccine effectiveness can be explained by HA mutations at epitope and/or surface sites, and whether changes affect amino acid function.
