Background

Genome-wide association studies (GWAS) have identified numerous disease-associated genetic variants over the past decades. However, the fact that over 90% of these variants are located in non-coding regions that do not encode proteins has remained a long-standing challenge. It has been difficult to clearly demonstrate which specific cells and regulatory factors are involved in non-coding variants causing disease.

Previous studies have primarily relied on bulk tissue analysis, which processes large numbers of cells simultaneously. This approach revealed limitations in failing to capture the unique regulatory signals of rare cell subtypes, instead diluting them within complex environments like the immune system. The practice of measuring chromatin accessibility (observing chromatin openness) and transcriptome expression levels in separate samples and then fitting them together via post-hoc statistics also had clear limitations. Using methods that indirectly connect data obtained from different cells imposed structural constraints on identifying the causal chain where a specific variant alters chromatin structure and directly disrupts target gene expression.

Key Findings

The researchers conducted single-cell multiome analysis on a large-scale population cohort, simultaneously measuring chromatin accessibility and the transcriptome within the same single cell nuclei. By decoding both the regions of open chromatin (scATAC-seq) and the actual transcribed gene expression levels (scRNA-seq) for each cell, they reconstructed the gene regulatory networks by cell subtype with high resolution.

The vast single-cell profiles collected provided decisive evidence linking the impact of genetic variants on chromatin structure (caQTL) with their impact on actual transcriptomic expression (eQTL). Analysis revealed that most disease risk variants selectively modify subtle chromatin accessibility sites where transcription factors bind. These modifications alter the physical interactions between distant promoters and enhancers located tens of kilobases (kb) or more away, thereby disrupting the expression levels of the target genes.

The impact of specific variants varied significantly depending on the type of immune cell. Non-coding variants associated with autoimmune diseases, such as rheumatoid arthritis and systemic lupus erythematosus, did not restructure chromatin across all immune cells but rather locally reorganized it only at very narrow cell differentiation stages, such as activated memory CD4+ T cells or specific B cell subtypes. This revealed that fine signals, which were dismissed as noise in previous bulk analyses, were actually the key molecular switches triggering disease onset.

Significance and Outlook

This achievement sets a milestone in converting genomic epidemiological data, which was previously limited to statistical associations, into actionable molecular biological targets. As it becomes possible to specify exactly which genes in which cells a specific disease variant targets, a foundation has been laid to significantly reduce the failure rate in the target validation stage of drug development. The development of precision therapeutics that selectively target cell-subtype-specific regulatory enhancers is also expected to gain momentum.

The task is also clear. Since this study was primarily conducted focusing on steady-state blood immune cells, it does not fully represent the dynamic regulatory changes occurring in actual tissue inflammation sites or lesion microenvironments. Further studies are required, including the expansion of cohorts to sufficiently capture genetic diversity among populations and the functional re-validation of the identified regulatory causal relationships in organoids or animal disease models.