I recently completed my PhD in the Computational Neural Data and Dynamics Lab with Dr. Eran Mukamel, where we were examining patterns of DNA methylation in cells from the brain. DNA methylation is a mechanism for regulating gene expression, and recent technological advances allow us to measure methylation patterns across the whole genome, as opposed to just a few positions. We were using this data to understand how methylation patterns change in the brain over development, how they differ across cell types, and also how these patterns are related to cognition and disorders such as schizophrenia.
The broader goal of my research is to utilize evolution as a form of computation in order to produce intelligent computational frameworks. This goal is built on several pillars. First, the human brain is merely a form of computation that can be reproduced as a computer simulation. Second, the computational framework of the human brain is special in the set of problems that it solves but not necessarily unique. Rather, there are likely many computational frameworks that can solve the same set of problems. Last, evolution is also a form of computation that is capable of building, or organizing, other computational frameworks such as the human brain. Consequently, we can simulate evolution on a computer to evolve computational frameworks that solve a specific set of problems. This approach serves as an alternative to designing the solutions by hand, which doesn’t seem to work well for large, complex behavioral spaces.