Exploring the molecular state space of cells in the brain
Speaker: Vilas Menon, Columbia University
Abstract: The brain contains a diverse array of cells with dramatically different electrical properties, morphologies, and connectivity patterns. Presumably, these differences are reflected in the underlying molecular composition of these cells. The advent of single-cell RNA-sequencing, which quantifies the gene expression-based molecular makeup of individual cells, has resulted in brain-specific data sets with measurements for thousands of features in each cell. In theory, each cell occupies a stable state in the overall molecular landscape, and cells of a given “type” are assumed to be close to each other in this landscape. Based on this assumption, the first major challenge in the single-cell RNA-seq field was to organize individual cells into coherent cell types based on the similarity of their molecular profiles. Approaches to solve this clustering/matrix decomposition problem have validated known cell type differences, while simultaneously providing new understanding into the organization of cells in the brain. In addition, insight from cell clustering leads to an intriguing new question: can existing information about the cellular landscape be used to infer the existence of stable, but hitherto unexamined, cell types? While not a mathematically rigorous problem, this question can nonetheless be approached with heuristic techniques to identify potentially novel cell types. Ultimately, developing systematic methods to predict the existence of unknown cell types has exciting future applications to the field of therapeutics and synthetic biology.
Room Reservation Information
Room Number: 106 McAllister
Time: 1:30pm - 2:30pm