CAREER: Data-driven Approaches for Investigating Olfactory System Heterogeneity

Bates College, Lewiston ME

Investigators

Abstract

The broad goal of this project is to develop and apply a suite of computational tools for studying the molecular and functional organization of the brain. To address whether circuits in a given brain area are organized as functionally diverse and heterogeneous modules despite apparent anatomical similarity, the PI and his associates will investigate the mouse olfactory bulb -- a brain structure dedicated to processing smell. Image data charting patterns of gene expression throughout the bulb will be obtained from open, digitally curated atlases (the Allen Brain Atlas (ABA)), and computationally mined to identify spatially structured motifs of gene expression. Electrophysiological recordings will also be made from slices of the bulb to directly test for the presence of organizational motifs identified in silico. These research activities will engage undergraduates extensively (the PI is a professor at a small liberal arts college), and provide them with experiences using computational approaches to study high-dimensional data sets. Additionally, the PI will promote the virtues of computational thinking and problem solving in biology through a redesign of his Introductory Neuroscience course. The new course will enlist students as practitioners in the analysis of data, and will make extensive use of open data sets from the ABA. A twin goal of the course redesign is to enhance diversity and retention in STEM disciplines through an emphasis on active, in-class learning. Of the myriad genes whose differential expression might demarcate distinct brain subregions, one can typically only test a small and idiosyncratic set in a single experiment, potentially missing important contributors to regional heterogeneity. Our first aim addresses this issue by analyzing the densely cataloged, whole-genome expression maps of the ABA to investigate the "zonal" molecular organization of the olfactory bulb. We will cluster gene expression profiles obtained from the ABA's in-situ-hybridization experiments to identify candidate spatial modes (e.g. "patchy", "periodic", "dorsomedial") of expression in the bulb. Our subsequent aims are centered around the collection of a densely sampled electrophysiological map of olfactory bulb principal neurons using in-vitro whole-cell slice physiology. Each recorded cell will be spatially registered to the same virtual bulb, and represented as a high-dimensional feature vector of physiological properties; these data will be clustered as above to test for physiologically distinct subregions, as well as for correspondences between "gene space" and "physiological space." We will use this same approach to directly and comprehensively investigate the degree to which sister and non-sister mitral cells (i.e. sharing a common glomerulus vs. not) are physiologically different.