Researchers have found that the DNA spools inside human cells are far less tightly wound than textbooks have long suggested. A study published in Nature applied an AI-driven analytical pipeline to single-molecule chromatin fiber data and identified 14 distinct structural states in nucleosomes, the protein-and-DNA units that package our genome. Rather than sitting in a simple binary of “open” or “closed,” most of these spools showed partial unwrapping or distortion, directly tied to how actively genes are read. An overview on phys.org emphasizes that this work overturns long-standing assumptions about how compact our DNA really is inside the nucleus.
Why Partly Unwound Nucleosomes Rewrite the Chromatin Playbook
For decades, chromatin biology operated on a two-state model: DNA was either accessible to the cell’s gene-reading machinery or locked away on tightly wound spools. That framework shaped how scientists interpreted gene regulation, disease mechanisms, and drug targets. The new findings replace that binary with a spectrum of 14 structural configurations, each associated with different levels of gene activity. The practical consequence is immediate. If most nucleosomes sit in intermediate, partly accessible states rather than sealed shut, then the signals controlling gene expression are far more varied than previous assays could detect.
The study used a method called IDLI, an analytical pipeline that processes data from single-molecule chromatin fibers captured through long-read methyltransferase footprinting. By marking which stretches of DNA around each nucleosome are exposed to an enzyme called EcoGII, the technique records a molecular imprint of how tightly or loosely each spool is wrapped at the moment of measurement. Older bulk assays averaged these signals across millions of cells, collapsing the diversity into a misleading binary picture. With IDLI, each individual chromatin fiber becomes a continuous record of nucleosome configurations along the genome.
This richer view of chromatin structure has direct implications for understanding how cells respond to signals. Instead of imagining that a gene is either locked down or fully open, the 14-state model suggests that promoters and enhancers can be tuned through subtle shifts in nucleosome distortion. For example, a transcription factor might favor a partially unwrapped configuration that exposes only a subset of its binding motifs, allowing graded rather than all-or-nothing responses to hormones, stress, or developmental cues. Such nuance could help explain why small changes in regulatory DNA sometimes lead to large differences in disease risk.
One testable extension of the work is whether these 14 states shift in predictable ways as cells age. If the IDLI pipeline were applied to matched datasets from young and senescent primary cells of the same tissue type, researchers could check whether certain nucleosome configurations become enriched or depleted with age, and whether those shifts correlate with increased transcriptional noise, the random fluctuations in gene expression that accumulate in older tissues. That experiment has not yet been reported, but the analytical tools now exist to run it, and the new structural catalog provides a framework for interpreting any age-related changes that emerge.
How the IDLI Pipeline and NN-HMM Identified 14 Nucleosome States
The core technical advance is a neural network paired with a hidden Markov model, referred to as NN-HMM. This system infers nucleosome footprints from EcoGII methylation patterns while correcting for enzyme access biases and sequencing errors that plagued earlier single-molecule approaches. The neural network learns to distinguish genuine nucleosome distortions from noise, and the hidden Markov model then segments each long DNA read into a sequence of structural states along the fiber. Together, they turn raw chemical marks into a map of how DNA is wrapped and distorted around histone proteins.
The data feeding this pipeline came from SAMOSA-style long-read methyltransferase footprinting, a technique in which the EcoGII enzyme is applied to intact chromatin inside cells, followed by PacBio single-molecule sequencing. PCR-free variants such as SMRT-Tag and SAMOSA-Tag, described in related methodological work, avoid amplification artifacts that can distort chromatin signals. Earlier reviews of single-molecule chromatin techniques, including work published in Frontiers in Cell and Developmental Biology, had already noted the promise of these approaches but lacked the computational framework to classify nucleosome conformations at this resolution. The IDLI pipeline fills that gap by integrating machine learning with biophysical modeling.
The result was a catalog of 14 distinct structural states. Some nucleosomes appeared nearly canonical, with DNA wrapped tightly around the histone core in the familiar 147-base-pair configuration. Others showed partial unwrapping at one or both entry and exit points, exposing stretches of DNA that would previously have been considered buried. Still others displayed internal distortions where the DNA bulged away from the protein surface without fully detaching, suggesting that bending and twisting forces can locally reshape the nucleosome without causing it to fall apart.
Quantitatively, the study reported that most nucleosomes fell into the partly accessible categories rather than the fully wrapped state, a finding that directly challenges the assumption that the default condition of chromatin is compaction. In regions of active transcription, the balance shifted even further toward distorted and partially unwrapped configurations, reinforcing the link between structural flexibility and gene expression. In contrast, more quiescent genomic regions retained a higher fraction of tightly wrapped nucleosomes but still showed a surprising diversity of intermediate states.
Transcription factors, the proteins that bind DNA to switch genes on or off, appear to actively reshape nucleosome structures rather than simply waiting for a spool to open on its own. The data showed that specific transcription factor binding events corresponded to specific distortion patterns, suggesting that the 14 states are not random thermal fluctuations but are programmed by the cell’s regulatory machinery. In some cases, binding footprints coincided with asymmetric unwrapping at just one end of the nucleosome, hinting that factors may pry open only the side they need to access. This mechanistic view aligns with biochemical studies that have long proposed that nucleosomes are dynamic, but it anchors those ideas in a concrete structural vocabulary.
Open Questions About Cell-Type Variation and Aging
The strongest gap in the current evidence is cell-type diversity. The primary datasets come from mammalian chromatin fibers, but the published work does not yet report whether the distribution of the 14 states differs systematically between, say, a liver cell and a neuron. If each tissue maintains a characteristic nucleosome state profile, that profile could serve as a new kind of epigenetic fingerprint, one with finer resolution than existing chromatin accessibility maps. Comparing such profiles across development could reveal when during differentiation particular distortion patterns become locked in.
A second unresolved question is reversibility. The institutional summary of the research raises the possibility that nucleosome distortions could be reversed, with implications for aging and disease. But no primary source material in the current reporting addresses whether these states are stable over time or whether they can be pharmacologically manipulated. If certain distorted configurations promote harmful gene expression programs in cancer or neurodegeneration, drugs that bias nucleosomes back toward less accessible states might have therapeutic value. Conversely, in contexts where beneficial genes are silenced, nudging nucleosomes into more open configurations could enhance expression without permanently altering the underlying DNA sequence.
Answering these questions will require longitudinal and perturbation studies that repeatedly profile the same cell populations under defined stimuli. Because IDLI and NN-HMM operate at single-molecule resolution, they are well suited to track how nucleosome states shift after exposure to signaling molecules, chromatin-modifying drugs, or environmental stress. Integrating those structural readouts with transcriptomic and proteomic data could ultimately reveal how the 14-state spectrum is wired into the broader regulatory network of the cell.
For now, the key message is that chromatin is neither simply open nor closed. It occupies a finely graded landscape of partially unwrapped and distorted nucleosomes, each configuration carrying information about the gene activity in its neighborhood. As methods like IDLI spread to more laboratories and more biological systems, that structural vocabulary is likely to become a central part of how researchers describe, and eventually manipulate, the genome in living cells.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
