Speaker: Dr. Debora Marks
Associate Professor
Department of Systems Biology
Harvard Medical School
Host: Dr. Michael Wiener
Prediction and design of proteins with neural machines
What can we do with billions of genomes and immune repertoire sequences? We now an amazing opportunity to develop machine learning methods that can exploit this enormous natural diversity to predict the effects of human genetic variation, to predict how viral genomes may evolve and to design biological sequences for therapeutics and biotechnology. I will demonstrate how unsupervised probabilistic generative modeling of sequences can give surprisingly direct answers to questions about 3D structures, dynamics, genetic variation and the design of biological molecules optimized for specific functions. I will end by introducing challenges for extending these methods to diverse applications to a broad range of biotechnology applications.
Event Categories: Archive, MPBP seminar