Convergent Functional Genomics of bipolar disorder: from animal model pharmacogenomics to human genetics and biomarkers

Abstract

Progress in understanding the genetic and neurobiological basis of bipolar disorder(s) has come from both human studies and animal model studies. Until recently, the lack of concerted integration between the two approaches has been hindering the pace of discovery, or more exactly, constituted a missed opportunity to accelerate our understanding of this complex and heterogeneous group of disorders. Our group has helped overcome this "lost in translation" barrier by developing an approach called convergent functional genomics (CFG). The approach integrates animal model gene expression data with human genetic linkage/association data, as well as human tissue (postmortem brain, blood) data. This Bayesian strategy for cross-validating findings extracts meaning from large datasets, and prioritizes candidate genes, pathways and mechanisms for subsequent targeted, hypothesis-driven research. The CFG approach may also be particularly useful for identification of blood biomarkers of the illness.

Figure 1

Initial Iteration of Convergent Functional Genomics

Figure 2

Figure 3

Expanded Convergent Functional Genomics Approach to Identifying Candidate Genes Involved in Bipolar Disorder and Related Disorders

Figure 4. Top candidate genes for bipolar disorder(s)

Probability pyramid generated by the tabulation of independent convergent lines of evidence (as shown in Figure 3B). Plain text-increased by methamphetamine. Italics-decreased by methamphetamine. Color coding-different target brain regions in our animal model. (a) from Ogden et al. 2004. (b) updated 2007 with new external lines of evidence published in the field since 2004. Once a gene is on the pyramid, it can only move up in priority as new evidence emerges over time. On the sides of the pyramids is depicted the scoring based on number of lines of evidence.

Figure 5

Expanded convergent Functional Genemonics (2007): Multiple Independent Lines of Evidence for Bayesian cross Validation