How does the tissue coordinate stochastic single-cell transcription into a spatially organized differentiation program?
Biology
Stem cell differentiation follows a deterministic sequence of gene expression changes, yet individual cells transcribe in stochastic bursts, leaving how tissue-scale order emerges from cell-level noise unresolved. Using intravital two-photon imaging of live mouse epidermis, I track transcription and protein dynamics to ask how differentiation is orchestrated through bursting kinetics within cells, communication among neighbors, and spatial localization of resulting transcripts and proteins. This work will provide a framework for understanding how tissue architecture shapes gene expression.
How can we interrogate the coupling between morphology and gene expression changes during differentiation?
Method
Fate decisions reshape both nucleus morphology and transcriptional state, yet the two are usually measured in separate experiments, leaving their coupling indirect. Using 3D nucleus segmentation and spot tracking from intravital imaging, I extract both readouts from the same cell, capturing how shape and transcription co-evolve as nuclei delaminate within intact tissue. The result is a single-cell level, spatiotemporal map of this coupling.
How do we organize complex 4D imaging datasets in the way that they stay traceable and reproducible?
Analysis
Intravital imaging accumulates large 4D datasets across animals, conditions, and timepoints, and analyzing these volumes carries substantial computational cost. Without a schema that holds up, a single broken metadata link can render a result untraceable and irreproducible. I design the data infrastructure: schemas, raw-to-derived linkages, HPC tracing, and self-contained HTML QC reports per experiment, so each new dataset slots into the project collection as a queryable object: directly comparable to any prior experiment, with full provenance from raw image to final figure. As the collection grows, so does what we can ask of it.