Gene Regulatory Dynamics in Human Brain

"The human brain comprises diverse cell types governed by distinct regulatory programs. While chromatin accessibility profiling has facilitated the identification of candidate regulatory elements, it provides only a limited, binary view of element activity. Here, we report joint single-cell multimodal profiling of transcriptomes and histone modification landscapes across 171 cell types from the human cerebral cortex, cerebellum, and pons. By integrating H3K27ac and H3K27me3 profiles with previous single cell maps of chromatin accessibility, DNA methylome and 3D genome architecture from the same donors, we reveal unprecedented details of heterogeneity in epigenome across brain regions and cell types. We utilized this integrative dataset to annotate both active and PRC2-mediated repressive chromatin states for over 500,000 candidate cis-regulatory elements and infer their putative target genes and transcription regulators. Furthermore, we construct cell-type-specific gene regulatory networks and compile a comprehensive catalog of candidate transcription factors that define and drive cell identity. Finally, we trained a deep learning model leveraging multi-dimensional epigenomic features to achieve high prediction accuracy for functional risk variants and enhancer activity. In summary, the single cell epigenome atlas of the human brain helps to resolve regulatory heterogeneity across brain cell types and provide a foundation for studying gene regulatory mechanisms underlying the heterogeneity of the human brain."

Data citation

No data citation listed

Funding agency

NIMH

Grant numbers

1RF1MH128838, UM1MH130994

Access

open

License

CC BY 4.0