Understanding gene regulation requires seeing it happen, in real time, in single cells, and in intact tissues. The imaging tools to do this at each scale have not existed. Our lab builds them.

Over the past decade, we have invented a series of interconnected imaging platforms that span three scales of biological organization:

Single-molecule methods. We developed new approaches for tracking and labeling individual proteins in living cells and whole organisms, including residence-time-resolved transcription factor imaging (Chen et al., Cell 2014) and stochastic protein labeling for long-term single-molecule imaging in vivo (Liu et al., PNAS 2018). These tools revealed how transcription factors search the genome, how molecular cargo is trafficked in neurons, and how disease-associated protein aggregates disrupt gene regulation.

3D ATAC-PALM. We invented a super-resolution imaging method that combines the Assay for Transposase-Accessible Chromatin with lattice light-sheet PALM microscopy to visualize nanoscale genome organization in single cells (Xie & Dong et al., Nature Methods 2020). Using 3D ATAC-PALM with genetic perturbations, we discovered that cohesin compartmentalizes the accessible genome and regulates gene co-expression patterns, not average expression levels, in single cells (Xie & Dong et al., Nature Genetics 2022; Dong et al., Nature Genetics 2024).

cycleHCR. This discovery motivated us to build a new platform for reading gene co-expression signatures in situ. cycleHCR is a deep-tissue spatial transcriptomics and proteomics method that images hundreds of RNA and protein targets with subcellular resolution in thick tissue volumes (>300 µm), with 20× higher throughput than prior multiplexed HCR-FISH approaches and a fully automated workflow (Gandin & Kim et al., Science 2025).

Each technology was built to solve a specific problem that the previous tool uncovered. This cycle of tool building and biological discovery drives everything we do.

Looking ahead, we are integrating these platforms with systematic genetic perturbations and machine-learning analysis to map the regulatory networks that control cell-fate decisions in mammalian development and disease. We design and build custom microscope hardware, automated fluidics, and control software in-house.