Organs have intricate spatial organization, whether it is the layers of the skin or the nephrons in the kidney. To study this spatial biology, researchers use spatial transcriptomic technologies that measure RNA molecules from cells and map them back to their location in a sample. But these methods typically construct the spatial transcriptome map by creating an image of the tissue, which can be expensive and time-consuming.
By leveraging barcode diffusion, Fei Chen developed an imaging-free spatial transcriptomic technology to unlock large-scale transcriptome positional activity.
Broad Institute of MIT and Harvard
Fei Chen and Chenlei Hu at the Broad Institute of MIT and Harvard have developed a new imaging-free spatial transcriptomics technology that tracks the diffusion of DNA barcodes between beads in an array to reconstruct the spatial organization of the tissue.1 In an interview with The Scientist, they described how they came up with this idea and their vision for a future with more scalable and accessible spatial transcriptomics.
What are the limitations of previous methods for spatial transcriptomics?
Chenlei Hu: One way that scientists perform spatial transcriptomics is by using sequencing to profile the RNA. Slide-seq, which we developed in our lab, is an example of such a method.2 It uses a spatial barcode conjugated to the RNA, enabling sequencing of all transcripts together. But we still need to use imaging to determine the position of those spatial barcodes, which requires advanced microscopes and takes time. Samples can only be up to three to five millimeters—even smaller than a mouse brain—so we often need to cut the tissue down to that size. Imaging limits the throughput of spatial transcriptomics, so we wanted to develop a system that no longer requires imaging.
What motivated you to design the new barcode diffusion-based approach?
Fei Chen: There is a theory that you can reconstruct spatial information from sequencing without ever having to image the sample. One advantage of this idea is that sequencing costs have been exponentially decreasing. Converting the spatial problem to a sequencing problem would make the method incredibly scalable and very easy for researchers to use.
Chenlei Hu leveraged barcode diffusion to develop an imaging-free spatial transcriptomic technology that can unlock large-scale transcriptome positional activity.
Irving Barrera
How do you reconstruct the spatial map of beads and RNA profiles in the tissue?
CH: Although it seems intuitive that scientists can use diffusion to infer spatial information about the beads, it is hard to actually do so. We started with physical diffusion models, but this was difficult because of the large amount of data and its noisiness.
FC: We spent a long time trying to figure out how to do it, and we kind of gave up for a while.
CH: One day I realized the data were similar to single-cell data. In single-cell analysis, researchers use dimensionality reduction to visualize the similarity between cells based on their gene expression. This is analogous to what we want in the spatial reconstruction problem: if two beads are always interacting, they should also be close together. We realized that we could just borrow a common dimensionality reduction method from single-cell analysis called uniform manifold approximation and projection.
What do you hope researchers achieve with this new method?
FC: The cost of collecting spatial transcriptomic data is very high right now, and this method makes it at least an order of magnitude cheaper. There are experiments, for example, in cancer, where scientists are currently collecting one representative section of each tumor. But they might gain more statistical power if they can sample the entire tumor.
We are collecting a lot of large human organs, mainly brains. We think we will be able to find cell type and gene expression patterns across really large length scales that we would never be able to observe otherwise—for example, across an entire human brain section.
Our technique is very scalable and easy for researchers to adopt. The cool thing is that we just send them beads, and they can follow the protocol. They do not need any special equipment. We want to lower the barrier to spatial transcriptomic analysis.
This interview has been condensed and edited for clarity.
- Hu C, et al. Scalable spatial transcriptomics through computational array reconstruction. Nat Biotech Nol. 2025.
- Rodriques SG, et al. Slide-seq: A scalable technology for measuring genome-wide expression at high spatial resolution. Science. 2019;363(6434):1463-1467.
