Single-Cell Atlas Connects the 3D Genome to Chromatin Changes

For the most part, every cell in the body has the same DNA, but each cell type, like brain or insulin-producing cells, uses it differently to perform specific functions. Epigenetic factors, which include chemical modifications to DNA and the physical 3D arrangement of DNA packed into the nucleus, help determine which genes are active and give each cell its unique function.

For years, single-cell technologies have allowed scientists to examine individual cells with remarkable clarity, but researchers usually examine genome architecture and epigenetic DNA markers as separate features. By studying these systems in isolation, and often in only a handful of tissues, scientists missed the delicate interplay between physical structure and chemical code inside a living cell.

Addressing this gap in a recent study published in Science, researchers at the Salk Institute for Biological Studies and their collaborators developed the first comprehensive single-cell atlas capturing two major epigenetic systems at the same time in the same cells: 3D genome folding and DNA methylation.¹ This atlas showed that physical DNA folding and chemical methylation marks vary independently across cell types, giving scientists an unprecedented reference to explore how non-coding genetic factors shape human health and disease.

In the human genome, gene regulatory sequences are often located far from the genes they control, which makes it difficult for researchers to identify which gene a regulatory region affects. By studying the 3D genomic architecture along with DNA methylation, scientists can better understand how these distant regulatory elements influence gene activity. As Jesse Dixon, a molecular biologist at the Salk Institute and study coauthor, explained, it’s like “trying to figure out the wiring diagram of a house when the lights in one room might turn on the lights in a completely different part of the house.”

To create their single-cell atlas, the researchers first isolated nuclei from over 86,000 single cells from 16 different human tissues, including the brain, heart, lungs, and liver. They then used single-nucleus methyl-3C sequencing to simultaneously analyze the 3D chromatin structure and DNA methylation patterns.

They found that different cell types organize their genomes in unique ways. Rather than following a single structural rule, some cells group active and inactive genomic regions into large compartments, while others use fine-scale loops to connect specific control switches to target genes. Many different cell types blend both approaches. Researchers had not fully appreciated this structural diversity across cell types before.

Looking at the atlas, Dixon and his team identified DNA regulatory elements like promoters and enhancers, alongside typical canonical methylation, where cytosines precede a guanine, and rare non-canonical methylation, in which cytosines are methylated without a following guanine. Mapping these rare patterns across individual cell types revealed a layer of genome regulation often missed in bulk tissue analyses. Ultimately, the team identified known genetic risk factors for conditions like atrial fibrillation, diabetes, and schizophrenia act within specific cell types to drive disease.

The researchers also found that the 3D DNA structure can change rapidly while DNA methylation changes more slowly. For example, mature muscle cells may already have their final structure even though their methylation patterns still look young. Identifying these epigenetic mismatches during cellular transitions allows researchers to expose the hidden molecular clocks driving disease progression, tissue regeneration, and cellular aging.

Thomas Gingeras, a geneticist at Cold Spring Harbor Laboratory who was not involved in the new research, praised the study, calling it a valuable resource. He added, “The most intriguing result is that these two dimensions do not always assign the same identity to a cell. That suggests that cell states are layered and that different regulatory systems may change on different time scales.”

The single-cell atlas serves as an invaluable reference blueprint for healthy gene regulation, laying the groundwork for future genomic discoveries. The dataset also addresses a critical bottleneck in the field by providing the high-resolution, cell-type-specific training data essential for building accurate predictive AI models of variant effects. To make this resource globally accessible, Dixon and his team launched an interactive web browser covering every human tissue and cell subtype profiled in the study. Reflecting on the atlas’s immense data depth, Dixon noted, “You could write a book instead of a paper.” Looking ahead, he and his colleagues plan to leverage this atlas to track how spatial genome organization alters throughout aging and disease pathogenesis.

“The atlas begins to define the organizational identity of genomic regions.” Gingeras added. “The next major challenge is to understand how regulatory molecules—particularly RNAs and their associated proteins—recognize and act upon that identity.”