Illuminating the Brain Architecture of Neuropsychiatric Risk: The Latest from PsychENCODE: Neuropsychiatric disorders have highly complex causes that involve hundreds of genes - a feature that hindered development of related treatments. Much of the genetic variation underlying neuropsychiatric diseases resides in noncoding regions of the genome, making establishing clear links between genetic variants and disease outcome especially challenging. Now, ten new reports from the PsychENCODE Consortium - a multidisciplinary effort established in 2015, and dedicated to illuminating the molecular mechanisms underlying schizophrenia, bipolar disorder and autism spectrum disorder - make progress in unraveling brain architecture relevant to these diseases. The studies are published in Science (7), Science Translational Medicine (2) and Science Advances (1), with highlights outlined below.
In a "flagship" study in Science analyzing brain development over time, Mingfeng Li and colleagues apply multiple genomic techniques to profile gene expression, epigenetic modifications and regulatory elements in different brain regions and cell types across the course of brain development. Critical expression differences tied to neuropsychiatric risk are most pronounced during early brain development and adolescence, they report. A second major finding is that neuropsychiatric risk genes converge on specific cell types and time points in brain development, providing new insights into when and where to study these disease mechanisms, and how to model them. A second flagship study, also in Science, integrates brain tissue and single-cell data from nearly 2,000 individuals with deep-learning approaches, the latter used to predict the risk of disease. In this study, Daifeng Wang and colleagues used single-cell sequencing to explain variations in basic cell types, such as excitatory and inhibitory neurons, with relevance to neuropsychiatric disease. The large scale of their data also allowed them to build molecular and regulatory networks linking noncoding variants of DNA to specific genes; they were then able to use these networks to link many more known psychiatric risk variants to genes than scientists have done. Finally, Wang et al. embedded the regulatory networks they'd built into an interpretable deep learning framework to predict the risk of neuropsychiatric diseases. Their predictions, they say, were much more successful than those relying on traditional genetics.
Two studies in Science Translational Medicine identify distinct patterns of genetic activity in the brain that are linked to schizophrenia and bipolar disorder, respectively. The findings may facilitate future drug discovery efforts for schizophrenia and bipolar disorder, which each affect tens of millions of people around the world and represent a massive economic burden - one analysis estimated that schizophrenia was responsible for over $155 billion in total costs in the U.S. in 2013.
A final study published in Science Advances by Suhn Rhie and colleagues is the first to produce a three-dimensional epigenomic map of a subset of neuronal cells that serve as a good model for epigenetic studies of the neurodevelopmental components of schizophrenia. The team's evaluation of these cells - taken from biopsies from living individuals with a diagnosis of schizophrenia, as well as from healthy controls - uncovered individual differences in epigenetic signatures at regulatory elements. The results may provide new targets for therapeutic intervention for schizophrenia, the authors say.
For these papers, the consortium developed analytical and biological tools that will be resources to the community. All data and associated analysis products are available from the consortium website (psychencode.org). They dedicate this series of papers to Pamela Sklar, one of the chief architects and leaders of the PsychENCODE Consortium.