Research Directions

Our lab works across quantitative imaging, molecular biology, and data science. We design custom optical instruments, develop molecular labeling and detection chemistries, and write the software that runs our microscopes and analyzes our data. We do this work because new biology requires new tools, and the most useful tools are designed for a specific unsolved problem, not in general.

Each of our major platforms (single-molecule tracking in live cells, 3D ATAC-PALM, and cycleHCR) was built because no existing method could answer the question in front of us. And each discovery a new tool made possible revealed the next question, which demanded the next tool. This cycle of invention and discovery is the organizing principle of the lab.

Our working hypothesis. We hypothesize that many genomic and epigenetic features, particularly long-range regulatory interactions, did not evolve to switch individual genes on and off, but to fine-tune the co-expression of multiple genes at the single-cell level. By analogy, these regulatory interactions act like the weights of a deep neural network, tuning the strength of gene-to-gene connections rather than the ON-OFF state of individual nodes. Together, these complex networks coordinate thousands of genes at their correct expression levels. The modulation that matters is often small in any single cell: a few percent change in transcript and protein abundance, or a relative shift in a high-dimensional gene-expression space. Yet collectively, these subtle effects are functionally important: they help drive cell-fate decisions and define each of us as individuals. At tissue scale, signals this small are difficult to study with current single-cell methods whose high noise floor can rival the biological effect. To read and study this kind of regulation at scale, we need precision measurements with low noise floors, high dynamic range, and the multiplexing capacity to track many regulatory variables in the same cells. This is why we build the tools we do: the precision of the measurement is itself the experiment.

At Johns Hopkins, we are building the next generation of tools in three areas:

Next-generation spatial omics. We are working toward higher gene multiplexing, faster acquisition, and compatibility with different types of tissues. This includes custom microscope builds, new fluidics architectures, new probe chemistry, and integration with expansion microscopy and tissue-clearing methods for whole-organ spatial profiling.

Multiscale integration: connecting genome structure to tissue maps. We are developing workflows that link quantitative measurements of chromatin organization in single cells with spatial transcriptomics in the same tissue. Combined with systematic genetic perturbations, including CRISPR knockouts, degron-mediated protein depletion, and intersectional genetic strategies, these integrated measurements will let us ask how specific regulators reshape chromatin architecture, alter spatial gene-expression patterns, and ultimately change cell fate within a tissue context.

Computational tools and AI. As our datasets grow in dimensionality (hundreds of protein + RNA targets × millions of cells × 3D volumes), we are building machine-learning assisted pipelines for automated cell segmentation, gene quantification, spatial clustering, sub-cellular structure characterization, and regulatory network inference.

We are committed to making our tools accessible to the broader research community. To this end, we keep optimizing our technology for scalability and ease of use for broad and robust deployment.