Large datasets of single-cell gene expression and cellular morphology provide an exciting opportunity to learn predictive models of cellular properties. Replicating the remarkable successes of generative AI models for vision and language in the cellular domain would be highly significant for biomedical science. In this talk, I will present three examples of how generative AI can predict key properties of cells, including their dynamics during differentiation, their responses to perturbation, and their morphological shapes.
Joshua Welch is an Associate Professor of Computational Medicine and Bioinformatics and Computer Science and Engineering at the University of Michigan. He earned his PhD in Computer Science from the University of North Carolina in 2017 and performed postdoctoral research at the Broad Institute of Harvard and MIT before starting at the University of Michigan in 2018. His team develops computational approaches for single-cell and spatial transcriptomic data analysis with applications to stem cell reprogramming and neuroscience. His work has been funded by the National Institutes of Health and the Chan Zuckerberg Initiative.