The Thomson Lab works across a variety of biological model systems to uncover new principles. Recent work has focused on systems including cytoskeletal active matter and cellular self-organization in neural circuit development. We apply tools from biochemistry, computer science, molecular biology, and physics. We are united in our goal to explore principles underlying collective organization and intelligence in biology and to reengineer biological systems ranging from cells to organisms.
Some of the most exciting frontiers of biological research involve understanding how cells manipulate matter in time and space to move, divide, and change shape. Our group is studying the self-organization of force and motion in cells by exploring defined “active matter” systems where cell-like behaviors can be achieved experimentally using a small number of purified proteins.
We are designing bio-inspired developmental algorithms that can be used to “grow” and self-organize neural networks and their architecture without human intervention. The ability to grow and self-organize artificial computational devices from a single unit will enable large scale generation of neuron-like computing architectures that are scalable, robust, power efficient, and reconfigurable.
Cell behavior is controlled by complex regulatory networks that allow single cells to respond to environmental and physiological conditions. Only recently have we had the experimental capacity to probe and perturb single cells with enough scale to parameterize full mathematical models of cellular regulation. We are developing machine learning methods that effectively automate the scientific methods to allow closed loop learning of regulatory network models through iterative loops of observation, model construction, prediction, and perturbation.