Super-resolution fluorescence microscopy, recognized by the 2014 Nobel Prize in Chemistry, particularly single-molecule localization microscopy (SMLM), has revolutionized the study of cellular ultrastructure by enabling nanometer-scale imaging of molecular positions. However, a fundamental limitation has persisted: while SMLM reveals where molecules are, it cannot simultaneously determine how they move. Resolving molecular dynamics requires sparse multi-frame trajectory tracking, making it difficult to capture spatial distributions and real-time motion concurrently. In live cells, the “static distribution” and “dynamic behavior” of molecules have long remained mutually exclusive, posing a major barrier to understanding dynamic cellular processes.

On April 28, 2026, the laboratory of Wulan Deng at the Biomedical Pioneering Innovation Center (BIOPIC), Peking University, published a research article in Nature Methods titled “Single-Molecule Localization and Diffusivity Microscopy Reveals Dynamic Biomolecular Organization in Living Cells.” This study introduces SMLDM, a novel live-cell single-molecule imaging technique that simultaneously resolves localization and diffusivity. By leveraging deep learning, SMLDM directly infers molecular trajectories, diffusion coefficients, and positions from a single snapshot, without relying on multi-frame trajectory linking. This establishes a new computational paradigm of “single-frame dynamic inference.” Compared with conventional tracking methods, SMLDM increases the density of dynamic single-molecule data by 50–300 fold, enabling the generation of high-density, super-resolved diffusion maps in living cells with single-molecule precision—thus achieving both spatial and dynamic measurements simultaneously. The method has already been successfully applied to multiple dynamic cellular processes.

This innovation arose from the team’s careful observation of experimental details combined with interdisciplinary thinking. In conventional live-cell single-molecule imaging, “sharp, tail-free” signals from fast imaging are considered ideal, while motion-induced fluorescence streaks are often treated as noise. The Deng lab challenged this assumption and recognized that these streaks encode the full motion of molecules during the exposure time. These often-overlooked “blurry traces” in fact contain critical information for decoding molecular dynamics.

To extract this information, the team developed an integrated computational framework combining deep learning and diffusion theory. Specifically, they constructed a convolutional encoder–decoder neural network, Deep-SnapTrack, trained on simulated dynamic single-molecule data to reconstruct “pseudo-track” from single short-exposure images. They further developed the TrackD algorithm to quantitatively estimate diffusion coefficients from pseudo-trajectory areas, and the TrackL algorithm to adaptively optimize localization precision across different dynamic regimes, surpassing traditional methods. Experimentally, by employing bright photoactivatable fluorophores and a dedicated U-Net model for high-density single-molecule segmentation, the team established a fully integrated pipeline—from image acquisition and molecule detection to trajectory reconstruction and kinetic parameter extraction—forming the methodological foundation of SMLDM (Fig. 1).

Figure 1. Overview of the SMLDM workflow and example data

By integrating SMLDM with PALM, the researchers further developed MPALM (mobility photoactivated localization microscopy) and applied it to four key biological contexts:

1. Chromatin spatial organization: Revealed low-diffusivity mesoscale domains in live-cell chromatin and uncovered a strong correlation between nucleosome mobility and chromatin density (Fig. 2).

2. Membrane receptor signaling: Characterized ligand-dependent biased clustering of μ-opioid receptors, providing a single-molecule dynamic perspective on GPCR signaling.

3. Focal adhesion remodeling: Elucidated the assembly, disassembly, and migration of focal adhesion nanoclusters, revealing spatial differences in molecular mobility between peripheral and internal regions.

4. Biomolecular phase separation: Captured early-stage condensation dynamics and revealed spatial heterogeneity in the maturation rates of micro-condensates.

Figure 2. MPALM analysis of histone H2B

Compared with classical Nobel Prize–winning techniques, the conceptual advance of SMLDM is both intuitive and profound: while conventional SMLM can only answer “where is the molecule?” from a single image, SMLDM can additionally determine “how it moves” and quantify its motion—all from the same snapshot. By adding adynamic dimension to spatial imaging, SMLDM represents a paradigm shift in live-cell single-molecule microscopy, from purely structural observation to integrated structural and dynamical measurements.

Dr. Zuhui Wang (postdoctoral fellow, Peking University) and Yiwen Liu (PhD student, BIOPIC) are co-first authors. Bo Wang (PKU–Tsinghua Joint Center for Life Sciences) is the second author. Dr. Xiangyu Liu (School of Pharmaceutical Sciences, Tsinghua University) contributed to the work. Dr. Wulan Deng is the corresponding author. The study benefited from guidance by Prof. Fan Bai, Hao Ge, Yujie Sun, and Xiaoliang Xie (Peking University), and Prof. Dong Li (Tsinghua University). This work was supported by the National Key R&D Program of China, the National Natural Science Foundation of China, the China Postdoctoral Science Foundation, the Beijing Advanced Innovation Center for Future Gene Diagnosis, the PKU–Tsinghua Joint Center for Life Sciences, and the State Key Laboratory of Gene Function and Regulation.

Original article:https://www.nature.com/articles/s41592-026-03078-x

Corresponding Author

Wulan Deng is a Principal Investigator at the Biomedical Pioneering Innovation Center (BIOPIC), the Beijing Advanced Innovation Center for Future Gene Diagnosis, and the PKU–Tsinghua Joint Center for Life Sciences, Peking University, and a Boya Young Scholar. She received her B.S. from Peking University in 2006 and her Ph.D. in Molecular and Cell Biology from the University of Pennsylvania in 2012. From 2013 to 2019, she conducted postdoctoral research at the Howard Hughes Medical Institute Janelia Research Campus and the University of California, Berkeley.

The Deng lab focuses on the intersection of super-resolution single-molecule live-cell imaging and transcriptional regulation. By integrating interdisciplinary approaches with advanced optical imaging, the lab has developed a multimodal quantitative imaging platform for in situ tracking and analysis of single-molecule dynamics in living cells. Their research spans three main areas: temporal regulation of transcription (particularly transcription factor–genome interactions), dynamic assembly of biomolecular complexes (including intrinsically disordered proteins) and 3D genome organization, and the development of cutting-edge live-cell imaging technologies. Guided by the principle that “biological questions drive methodological innovation, and new methods enable biological discovery,” their work aims to provide new visualization frameworks for understanding transcriptional regulation in stem cell differentiation and disease.