BEARS 2019: Jennifer Listgarten

Machine learning for protein engineering

Berkeley Annual Research Symposium (BEARS) 2019

Jennifer Listgarten

Prof. Jennifer Listgarten in front of part of a 2D visualization of a machine-learning-extracted "latent" space which converts triplets of nucleotides (codons) into real-valued vectors--- part of her process in silico protein engineering.


With the advent of more and more high-throughput technologies to measure protein properties of interest such as binding, expression, fluorescence, the time for machine learning to act synergistically with protein design is here. I will touch on two stories in this space. The first will be about how machine learning can be leveraged to improve CRISPR gene editing. The second will touch on how one can accelerate design/optimization of proteins with machine learning approaches--- a sort of in silico approach to the method of Directed Evolution, which won the 2018 Nobel prize in Chemistry.


Jennifer Listgarten is a Professor in the EECS department and Center for Computational Biology, a member of the steering committee for the Berkeley AI Research (BAIR) Lab, and a Chan Zuckerberg investigator.  From 2007 to 2017 she was at Microsoft Research and before that, earned a PhD in the machine learning group at the University of Toronto.  Her expertise is in machine learning, applied statistics and computational biology. She is interested in both methods development as well as application of methods to enable new insight into basic biology and medicine.  Current areas of interest include: computational methods for protein design/engineering for properties such as expression, flurorescence, binding, stability, etc.; similar methods applied to molecule design; drug repositioning and discovery; machine learning methods development, and in particular at the intersection of graphical models, neural networks and variational inference, as well as inverting black box probablistic functions to perform input optimization of probabilistic functions; genetic association studies with complex, high-dimensional traits such as image volumes over time.