From Genomic Association to Causation: A Convergent Neuroscience Approach 2.0

Presenter:

Linda Brady, Ph.D.
Division of Neuroscience and Basic Behavioral Science

Goal:

This re-issued initiative aims to support studies that leverage the large and diverse datasets being generated across levels of analysis (e.g., genetic, neurobiological, clinical) in order to develop novel theoretical frameworks for multiscale modeling and in silico experiments. These studies are expected to be designed to generate testable predictions for wet-lab experiments specifically to validate or invalidate causal relationships across these levels of analysis. The goal is to understand the biological mechanisms by which disease risk translates to disease pathophysiology, with potential for the unbiased identification of novel targets for therapeutic intervention.

Rationale:

Advances in informatics and new technologies enabling unbiased, automated, and scalable approaches, paired with collaborative team science, are leading to the generation of large complex datasets. This includes data identifying robust genetic associations with mental illnesses, high resolution transcriptomic, proteomic and higher order functional cell phenotyping, micro- and macro-connectomics, automated measurements of naturalistic behaviors and task performance, and rich clinical datasets including electronic medical records. Despite these advancements at each level of analysis, a biological understanding of mental illnesses remains elusive. This is in part due to the multifactorial contributions to etiology, heterogeneity in disease course, and the likelihood that the causal factors and their pathophysiological outcomes are highly distributed across genes, molecules, cells, circuits, and cognitive domains of function. Most studies that cross levels of analysis are mainly correlative or select candidate components in order to demonstrate causality. Decoding the multidimensionality of mental illness remains a ‘hard problem’ in neuroscience.

The convergent neuroscience approach is designed to allow biomedical researchers to further explore large and diverse datasets arising from clinical, genetic, and neurobiology research in collaboration with experts from orthogonal disciplines (e.g., mathematics, computation, physics, engineering). Applicants will be encouraged to mine the wealth of unbiased, large scale data through multi-scale modeling and related computational methods such as machine learning. The purpose of such models is to make realistic and testable scientific predictions that incorporate robust genetic and/or environmental risk factors, account for brain and behavioral complexity, and address problems related to mental illness mechanisms, developmental trajectory, prevention, diagnosis and/or treatment among individuals, groups and within populations. Resulting predictions shall be tested by experimentation in order to mechanistically explain mental illnesses pathophysiological features by linking genetic, molecular, cellular, circuit, and behavioral processes. Several convergent neuroscience-solicited topic areas have not yet been supported from the initial Funding Opportunity Announcement and would benefit from the CN structure that this re-issued initiative would enable.