Researchers at ODU Shed Light on Brain Circuitry
By Jim Raper
Shuiwang Ji, an assistant professor of computer science at Old Dominion, leads a research team that has found a reliable and surprisingly simple way to predict sequential functions of the mammalian brain based on gene expression patterns.
This research focuses on understanding the connectome of a brain, which is a circuit of neural connections that plays a pivotal role in generating the cognition, emotion and perception of an organism. Neurological diseases, such as autism and schizophrenia, are commonly found to be associated with abnormal brain connectivity.
Findings of Ji and his collaborators were published online by the journal NeuroImage in September and will be included later in the print version of the journal. The title of the article is "Integrative Analysis of the Connectivity and Gene Expression Atlases in the Mouse Brain" and the authors also include Ahmed Fakhry, an ODU graduate student, and Houtao Deng, who works for the private company Intuit in Mountain View, Calif.
The researchers used a database called the Allen Mouse Brain Atlas compiled by the Allen Institute for Brain Science in Seattle, which is an independent, nonprofit medical research organization. This online resource provides a three-dimensional map of gene activity throughout the adult mouse brain, showing which genes are turned on where. It has become a resource for thousands of brain researchers around the world.
The mouse brain has organization and basic parts similar to other mammalian brains, and studies of it are often designed to shed light on functions of the human brain.
Ji's efforts in the past several years have been to turn the "snapshots" of brain activity that the atlas provides into a "moving picture" of connectivity.
Brain function is the result of interneuron signal transmission controlled by the fundamental biochemistry of each neuron, which is determined by spatiotemporal gene expression and regulation encoded into the genomic regulatory networks.
Much is known about this system, but the information is mostly static. "Systematic studies of the relationship between gene expression and connectivity in a mammalian brain are lacking," Ji said. "We employ computational models for predicting brain connectivity from gene expression data."
When the researchers analyzed the expression patterns of 4,084 genes, they obtained a predictive accuracy of 93 percent for how that expression translates into connectivity signals. More importantly, they found that some genes are more involved than others in connectivity-related processes and that by analyzing the expression of only 25 "highly ranked" genes they could predict connectivity with 91 percent accuracy.
Ji is the lead investigator at present on a $600,000 project funded by the National Science Foundation (NSF) to explore the mapping of the intricate workings of the mammalian brain. The grant is from NSF's Directorate for Biological Sciences, Division of Biological Infrastructure.
Ji, who leads the Computational and Biological Learning Laboratory at ODU, is collaborating on the project with: Patric Lundberg, associate professor in the Department of Microbiology and Molecular Cell Biology at Eastern Virginia Medical School; Axel Visel, staff scientist of the Genomics Division of Lawrence Berkeley National Laboratory in California; and Michael Hawrylycz, senior director for data analysis and annotation for the Allen Institute.
Ji and his collaborators said when they started work on the project that they planned to develop novel computational methods for analyzing the Allen Mouse Brain Atlas data set, with the goal of delivering a global (brain-wide and genome-wide) analysis that would elucidate the anatomic and genetic landscapes of the mammalian brain.
Although the Allen Mouse Brain Atlas contains voluminous data, charts and 3-D images showing where certain genes signal specific functions, not much is known about how the full neurological system operates within a timeframe. Models and simulations available to researchers today might be compared to a series of snapshots. Ji wants to make it possible for the data to be presented in an uninterrupted, full-brain simulation - a moving picture in 3-D, if you will.
The problem he and his team face is the sheer volume of gene-triggered connections within each mammalian brain - many millions of them - that work in concert to drive reactions and interactions relative to outside stimuli.
The researchers have developed tensor computation techniques to allow for data compression and visualization. This is a highly sophisticated means of sampling a lot of diverse data in order to come up with coherent patterns and reliable predictions.
Their analysis, the researchers believe, has the potential to enable and accelerate biological insights and generate experimentally testable hypotheses.
According to the project summary: "In particular, this project is expected to enable significant progress in the areas of (1) understanding the structure and development of the mammalian brain through the multidimensional analysis of genetic networks that can be identified through systems-level analysis of gene expression, (2) understanding molecular pathways that are active and contribute to the function of the mammalian brain, and (3) understanding how disruptions of genes and thus molecular pathways might contribute to pathologies of mouse and, ultimately, human brains.