Meet Our Researchers


Postdoctoral Fellow, Macosko Lab, Broad Institute of MIT and Harvard
The Question

Thanks to tools like single-cell technologies, we are generating gargantuan amounts of data about brain cells. For example, one experiment conducted by my lab examined 690,000 cells from nine regions of the adult mouse brain, identified 565 distinct groups of cells, and then compared and contrasted all the cell types across regions of the brain. As you might imagine, this took quite a while! It made me wonder: Could we develop a computational technique that would allow us to analyze brain cell data more rapidly?

The Approach

Borrowing from the world of pure mathematics, my colleagues and I will use our BroadIgnite award to develop a new data-analysis model that can streamline our efforts to uncover biologically significant representations of brain cell data. This model, which we call a nonlinear dimensionality approach, will employ machine learning. If it works, it could become the basis of a software tool that labs all over the world could use to gain insights into the roles brain cells play in schizophrenia, autism, and Alzheimer’s disease.