Project Description:
This project uses deep learning to analyze electron microscopy (EM) images of Mouse Hepatitis Virus (MHV), a model for coronaviruses, focusing on the structural difference between wild type and mutant variants. Manual segmentation was used to establish ground truth masks, which were then used to train a U-Net model for automatic segmentation as well as a ResNet18 classification model to distinguish the variants. Finally, morphological analysis of spike and body regions revealed consistent differences between the wild type and mutant strains. This pipeline lays the groundwork for future efforts to automatically annotate structural differences in viral morphology, paving the way for studying structure-function relationships in virology, immunology and pathology.
