What can we do with a million or a billion genomes? Understanding how variation across genomes shapes the properties of biomolecules, cells, and organisms is a foundational question in biology. Professor Marks will demonstrate how probabilistic generative modeling of genetic variation can give surprisingly direct answers to questions about 3D structures, dynamics, the effects of mutations and the design of biological systems. Their new work extends from the undirected models of genetic variation to deep directed models using newly developed variational autoencoders, and autoregressive models that do not depend on sequence alignments. From purely unsupervised learning, the Marks group has improved on prior-art for predicting the effects of mutations and successfully designed optimized antibody libraries. Professor Marks will introduce challenges for extending these methods to diverse biomedical and engineering applications, with specific examples of successful probabilistic models to generate novel functional biotherapeutics.