Chinese Immune Multi-Omics Atlas.

Human peripheral blood exhibits molecular and cellular heterogeneity across populations, yet the underlying mechanisms remain unclear. We present the Chinese Immune Multi-Omics Atlas (CIMA), characterizing molecular variations linked to sex, age, and genetic variants through multi-omics analysis of more than 10 million circulating immune cells from 428 Chinese adults. CIMA established an enhancer-driven gene regulatory network comprising 237 robust regulons; identified 9600 eGenes and 52,361 caPeaks at cell type resolution; and revealed pleiotropic associations among immune-related disease risk loci, cis-expression quantitative trait loci (QTLs), and chromatin accessibility QTLs. Furthermore, the cell language model CIMA-CLM predicted chromatin accessibility and evaluated the effects of noncoding variants from chromatin sequences and gene expression. CIMA provides a comprehensive reference for immune-related disease research. Editor's summary: Single-cell assays can give increased resolution over bulk experiments, but assessing multiple-omics in the same individuals, and particularly in individuals not of recent European ancestry, is still uncommon. Yin et al. performed single-cell sequencing assays for RNA levels (scRNA-seq) and chromatin accessibility (scATAC-seq) on immune cells from 428 healthy Chinese individuals, incorporating genotyping, lipids, and other biochemical markers. The authors assessed the impacts of age and sex on these outcomes, integrated transcription and chromatin data to identify enhancer-gene networks, and identified genetic variants associated with gene expression and chromatin accessibility. This atlas will serve as a resource for groups looking to better understand immune cell biology. —Corinne Simonti INTRODUCTION: Understanding the genetic regulatory mechanisms that drive human immune variation is critical for elucidating susceptibility to immune-mediated diseases. The advent of single-cell genomics provides the resolution necessary to map these genetic effects on gene regulation, resolving how they function in a cell type– and context-specific manner across the full continuum of the immune system. RATIONALE: A mechanistic understanding of immune variation and immune-mediated disease requires integrating transcriptomic data with epigenomic profiles to capture the effect of noncoding variants on chromatin. Progress has been impeded by the scarcity of large-scale single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) datasets from peripheral blood mononuclear cells (PBMCs) and the bias of existing resources toward individuals of European ancestry. We constructed the Chinese Immune Multi-Omics Atlas (CIMA) to bridge these gaps, providing a foundational resource to dissect the regulatory architecture of the immune system in the Chinese population. RESULTS: CIMA was developed by profiling 10,247,216 PBMCs from 428 Chinese adults using single-cell RNA sequencing (scRNA-seq) and scATAC-seq. Iterative clustering and hierarchical annotation of these cells defined 73 transcriptionally distinct immune cell types, enabling a systematic analysis of molecular variations linked to sex and aging. To uncover regulatory mechanisms, we mapped 338,036 candidate cis-regulatory elements (cCREs) using scATAC-seq. By integrating these chromatin maps with scRNA-seq data, we constructed enhancer-driven gene regulatory networks (GRNs) connecting 84,625 regulatory regions to 13,645 target genes. Furthermore, we identified cell type–specific and age-associated GRNs, revealing key transcription factors (TFs) in immune cells. By combining single-cell data with whole-genome sequencing (WGS) data, we performed cell type–resolved quantitative trait locus (xQTL) mapping and identified 9600 eGenes and 52,361 caPeaks. We also detected dynamic expression quantitative trait loci (eQTLs) along developmental trajectories in monocytes and B cells. Integrating these xQTL results with genome-wide association study (GWAS) summary statistics, we identified 1196 significant summary data–based Mendelian randomization (SMR) associations across 68 immune cell types; 73.2% of these associations were significant in only a single cell type. We revealed cell type–specific and shared pleiotropic associations that link genetic variants to chromatin accessibility, gene expression, inflammation-related circulating proteins, and disease risk. For instance, the variant rs34415530 exhibited pleiotropic effects, influencing both circulating interleukin-12B (IL-12B) protein levels and asthma susceptibility by mediating its effect on IKZF4 expression specifically in CD4+ FOXP3+ regulatory T cells. Finally, we developed CIMA-CLM, a cell type–specific language model integrating chromatin sequences and scRNA-seq data. The model accurately predicts chromatin accessibility with high consistency with experimental peaks across cell types. Furthermore, in silico mutagenesis confirmed its utility in assessing noncoding variant effects. CONCLUSION: CIMA provides a comprehensive population-scale immune multi-omics resource that resolves the cell type–specific regulatory architecture of the immune system. Our work offers a framework to refine our understanding of human immune diversity and dissect the genetic basis of immune-mediated diseases. CIMA: A comprehensive immune multi-omics atlas of a Chinese population cohort. More than 10 million circulating immune cells from 428 Chinese adults were profiled using multi-omics and were elucidated through constructed GRNs. Cell type–resolved xQTL mapping pinpointed thousands of eGenes and caPeaks. SMR integration with GWAS data revealed cell type–specific associations linking genetic variants to immune-mediated diseases. A cell language model was further built to predict chromatin accessibility. ILC, innate lymphoid cell; HSPC, hematopoietic stem and progenitor cell; TF, transcription factor; Treg, regulatory T cell; DC, dendritic cell. [ABSTRACT FROM AUTHOR]