DNA Microscopy: Seeing by Sequencing Revolution!

Sequencing has changed how we think about biology, medicine and even the environment: can it even change how we image things? Today we meet DNA microscopy, able to image samples in 3D using DNA.

Also, we have reached (and passed) 100 subscribers! Thank you for being part of it!
Every issue of Plenty of Room is a step toward building a space for interesting science, without the paper pile. I’m glad you’re here!

Let’s dive right in.

Seeing by Sequencing: What If You Didn’t Need a Microscope?

What is Spatial Transcriptomics?

Spatial transcriptomic is the hottest topic in biology (okay maybe not but it is popular). Researchers want to know not only what genes are expressed in a cell or tissue, but also where they are. In traditional sequencing techniques, all the cells in a tissue are destroyed at the same time, so their spatial organization is lost.

But information about the spatial context matters, because it:

• Reveals tissue structure: Links molecular changes to histology, helping us understand development, disease progression, and tissue architecture.

• Uncovers rare niches: Pinpoints small, spatially confined cell populations (for example tumors or stem‐cell niches) that bulk methods miss.

• Informs therapies: Spatial patterns of immune and tumor cells can predict treatment responses and guide precision medicine.

Spatial transcriptomics keeps this crucial information, allowing you to dive deep into it!

Different Flavors of Spatial Transcriptomics

Many techniques are available for spatial transcriptomics, and they have different balances of spatial resolution, scalability and gene coverage. They can be divided in a few categories:

1. Imaging-Based (Targeted): Uses probes to read out RNA in situ, one spot at a time. Examples are FISH or in situ sequencing. These techniques have high resolution, but require a priori knowledge of the RNA you are looking for, limiting their use to only known samples

2. Capture-Array and Bead-Based (Untargeted): Transfers mRNA from a whole tissue section onto a barcoded surfaces, then sequence everything. These techniques (10x Visium, Slide-seq, and more) are powerful, but they impose a fixed 2D grid and trade off resolution versus coverage.

But what if you could image gene expression in 3D, without a microscope or prior knowledge?

The DNA Microscopy Promise

DNA microscopy is a sequencing-only, optics-free way to “see” your biological sample. It encodes spatial relationships directly into synthetic DNA molecules, then it reconstructs images just by sequencing.

This sounds very cool:

• No slicing: It can (in theory) image whole, 3D tissues intact

• No labels: It doesn’t require staining or pre-selected targets

• Sequence-driven: It reveals gene identity and position in one go

But so far, it’s practical applications has been limited:

• Limited to 2D

• Poor resolution

• Limited computational scalability to large samples

This brings us to today’s paper (you can read it here!). In this work, the authors extend DNA microscopy to 3D imaging of whole zebrafish embryos. Crazy! They captured both the identity of genes and the spatial context: all with a single measurement.

So, how does it work?

3D DNA Microscopy: The Inner Workings

Okay, so there is a lot of DNA involved here, but bear with me, because it’s pretty cool.

TL;DR: DNA microscopy layers short-range, anchored DNA nanoballs with long range, unanchored RNA “probes” to create a DNA network. This network is then analyzed and used to create a micrometer-scale 3D map of gene expression, using only sequencing!

Now, here is how it works more in detail:

1. In situ barcoding of transcripts

   Inside a fixed embryo, every RNA is reveres-transcribed and ligated to a unique molecular identifier (UMI). This short, random DNA sequence tags the transcript, linking the sequence identity back to the spatial position.

2. Building local “nanoballs” for fine details

   Each circular UMI is amplified by rolling-circle amplification (RCA), creating a DNA nanoball (technical term, I swear) of around 1 µm. When those of these nanoball sit next to each other, a short DNA “bridge” (a UEI, unique event identifier) forms, capturing sub-micron proximities.

3. Capturing longer-range links

   The sample is embedded in a reversible PEG hydrogel and subjected to in vitro transcription (IVT), which turns DNA into RNA. This RNA can drift tens of microns, before occasionally recombining back into DNA bridges. These event add “longer range” UEIs, ensuring the global network of tags stays connected and we don’t lose large-scale shape.

4. Sequencing and data assembly

   Okay, the experimental part is done! Sequencing is the next step. We want to get two things: the UMI’s gene insert (to know which gene it came from), and the UMI-UEI-UMI bridges (so we count how often each pair of UMIs linked). In this way, we can create a list of barcodes and their gene identities, and a list of the bridges between them.

5. Reconstructing 3D position

   I will not pretend that I understood this part. I didn’t: I got the general gist of it though! And it’s very cool (don’t forget to read it here). So, the authors assume that high linkage counts imply close proximity between two points, low counts imply a greater distance. Using a geodesic spectral embedding (GSE) approach, the algorithm handles both local (nanoball) and global (hydrogel) scales, turning linkage counts into x-y-z coordinates for each UMI. It’s like Google Maps for sequencing!

Validation: Did it Work?

So, this is most of the paper: developing this new cool 3D imaging technique. But they also compared it to Stereo-seq, an established array-based spatial transcriptomic technique. Here, they used 2D zebrafish embryo sections, testing:

• Head vs. tail gene panels: They used Stereo-seq to pick “type A” (anterior-enriched) and “type B” (posterior-enriched) gene sets based on their spatial positioning in the zebrafish sections.

• Gradient concordance: Mapping the two types of genes in both Stereo-seq spots and DNA-microscopy UMIs revealed nearly identical anterior–posterior expression gradients, confirming that 3D DNA microscopy accurately captures true spatial patterns!

Key Takeaways

Cool new technique! They solved two of the limitations of DNA microscopy:

• It now works in 3D

• It scales to large samples like whole embryos

This makes it much more useful. In addition, sequencing have gotten cheaper and faster, making it possible to expand its uses.

Of course, there are still limitations:

• Resolution is still around 1 µm, so you can’t see single cells.

• Transcript capture is still relatively low

• The technique can be expanded, for example using antibodies for proteomics studies!

But this was a very cool work: go and read it for yourself here!

If you made it this far, thank you! Can you see yourself getting rid of your microscope? What else could we couple to sequencing? Have a new idea for a paper for me to cover? Reply and let me know!

More Room:

• Walking to the Target: DNA sensors are pretty cool (I work on them myself these days). But this paper adds a twist: DNA walkers. It introduces an enzyme-free 3D DNA walker with dual capture and output, enabling ultrasensitive SNP detection. Powered by ligase chain reaction (LCR) and entropy-driven amplification, it detects mutations at ~30 aM and distinguishes 0.01% mutation frequencies. Validated on soybean genomes, this system offers a powerful tool for precise, enzyme-free genotyping.

• From 0D to 2D: Apparently, there is such a thing has zero dimensions: and apparently it’s DNA. This work presents a DNA-guided method to assemble 0D–2D (heterostructures metal sulfide nanoparticles on MoS₂) with precise nanoscale control. The approach enables non-destructive tuning of optoelectronic properties and wavelength-specific photoresponses, offering high sensitivity and scalability for future optoelectronic devices. Very cool! I am getting more and more interested in this stuff.

• Cancer, Apoptosis and 3D Models: Ah, good old DNA origami to kill cancer! It makes me nostalgic. But this study goes deep, developing design principles for DNA origami-based cancer therapeutics targeting the Fas receptor to induce apoptosis. Using 3D cancer spheroids, the researchers show that nanoparticle penetration depends on DNA origami size, and therapeutic effectiveness is driven by how FasL is attached. Optimized nanoagents fully eliminated cancer cells in the spheroids, highlighting key factors for designing DNA nanostructures for tumor environments.

• Know someone who’d love this?
  Pass it on! Sharing is the easiest way to support the newsletter and spark new ideas in your circle.

• Got a tip, paper, or topic you want me to cover?
  I’d love to hear from you! Just reply to this email or reach out on social.