The lectures on protein design below are part of a summer workshop series on synthetic biology organized by he Canadian Synthetic Biology Education Research Group CSBERG.
Methods in Computational Protein Design
Summary: An overview of the methods available for computational protein design. We focus on the intuition behind the docking, MD simulations, Rosetta Design, machine learning models, transfer learning, and active learning. The point is to encourage computational thinking when thinking about design problems in synthetic biology.
Machine Learning for Protein Design
Summary: An introduction to the mathematics underlying the common machine learning algorithms used for protein design. We cover neural networks, gradient descent, loss functions, protein representations, Bayes rule, maximum likelihood estimate, MAP estimate, an application of variational Bayes to protein design, and generative models. The main point is how flexible machine learning approaches to protein design can be.
Summary: A demonstration of the common commands used in VMD to visualize protein systems. We look at PETase (plastic degrading enzyme) interacting with a plastic polymer and the SARS-CoV2 spike protein interacting with the ACE2 receptor.