Spatial Transcriptomics

Cells do more than control how much RNA they make—they also control where it goes. An RNA’s location can determine where its protein product is made and which molecular partners it encounters. We develop chemical tools to map RNAs within living cells and understand how their locations change during stress, cell division, and other cellular events.

Our light-activated RNA labeling method, CAP-seq, uses a targeted photosensitizer to chemically tag nearby RNAs, which we then enrich and identify by sequencing. We first used CAP-seq to map RNAs near the endoplasmic reticulum and mitochondria, revealing enrichment of transcripts encoding ribosomal and oxidative phosphorylation proteins at the outer mitochondrial membrane (Nat. Chem. Biol., 2019). We also extended this chemistry to DNA, allowing us to map chromosome regions near the nuclear lamina without formaldehyde crosslinking (Angew. Chem. Int. Ed. Engl., 2020).

We have since used CAP-seq to follow how RNAs gather and reorganize under stress. By examining stress granules as they assemble and disassemble, and targeting different proteins within them, we found that their RNA contents are dynamic and unevenly distributed (Nat. Commun., 2023; Anal. Chem., 2025). More recently, directed evolution produced Lantern, a photocatalyst that labels nearby RNAs and proteins in less than a minute. This faster labeling helped us examine early stress-granule assembly, including the recruitment of RNAs rich in the modification m⁶A (Nat. Chem. Biol., 2026).

In parallel, we develop enzyme-mediated RNA labeling. Introducing biotin-aniline as an APEX2 substrate enabled efficient labeling of nearby RNAs; combining this approach with metabolic RNA labeling led to MERR APEX-seq, which selectively profiles newly transcribed RNAs at specific cellular locations (Angew. Chem. Int. Ed. Engl., 2019; Cell Chem. Biol., 2022). We recently evolved an APEX2 variant with approximately twice the RNA-labeling efficiency while retaining spatial specificity. Using it, we identified ANLN mRNA near the midbody during cell division and found that its localization depends on the emerging protein chain (Adv. Sci., 2026).

Knowing where an RNA sits is only part of the story—we also want to know where it is being translated. Our optoRibo-seq method combines light-activated proximity labeling with ribosome profiling to capture local translation with one-minute temporal resolution. Applied at the endoplasmic reticulum, it revealed changes in the local translation of transcripts involved in protein folding and targeting during stress (PNAS, 2026).