Machine Learning Approaches and Methods for Discriminating B-Cell Epitopes, specifically in COVID-19 Spike Proteins

Abstract

The creation of physical protein models would be useful in a variety of different situations, from helping students who are visually impaired to providing a new perspective on how molecular structures that normally cannot be seen without the use of specialty equipment. Through the process of 3D printing, a digital model can be converted into a physical object, enabling humans to interact with this form of data more efficiently. Protein structures were constructed in a 3D modeling program (OpenSCAD) by assigning primitive shapes such as spheres and cylinders to coordinate information obtained from the Protein Data Bank (PDB). A Python script was used to convert the data into a code format accepted by OpenSCAD. From there, the model could be rendered and then printed using a variety of different printing mediums. Through this process, aids could be created to help teach a variety of classes. Improvements to this program, such as being able to process other types of molecular data for the creation of more models, could be done in the future.