Layers make things a lot more interesting—a cake with chocolate, vanilla and strawberry layers is much more decadent than a single-tier vanilla one. And now the labs of Rong Fan (Yale University), Gonçalo Castelo-Branco (Karolinska Institute), and Yanxlang Deng (now at University of Pennsylvania) have put the icing on the cake with a new paper describing how to get multi-layered spatial epigenomic and transcriptomics info at near-single-cell resolution.
Single-cell multiomics techniques can tell you a lot about omics layers, but they don’t pinpoint where things are happening. Recent spatial epigenomics and transcriptomics methods provide locations, but only give info on a one tier at a time.
So Fan’s team started building a more complex confection and as an initial step, developed DBiT-seq, combining high-spatial-resolution sequencing and deterministic barcoding, mining cells for both transcriptomics and proteomics localization data. Then, with Castelo-Branco’s lab, they developed spatial ATAC-seq, combining Tn5 transposition activity with barcoding.
Now, continuing on that theme, the collaborators added another set of strata: epigenomic and transcriptomic co-profiling.
The researchers let you choose the flavor of the epigenomics layer. For chromatin accessibility info, use spatial ATAC-RNA-seq. The first step is:
- Add a Tn5 transposition complex and a DNA adaptor/universal ligation linker to a piece of tissue, along with a biotinylated DNA adaptor/universal ligation linker with a poly-T sequence that can help start reverse transcription later on.
But if you prefer histone mod data in your tier instead, then use spatial CUT&Tag-RNA-seq. The first step here is:
- Add an antibody against specific histone mods (H3K27me3, H3K27ac, or H3K4me3) to the tissue and this time, use protein A-tethered Tn5-DNA.
Then, with either method:
- Place a microfluidic chip on the tissue to introduce barcodes in one direction that ligate to the universal ligation linker
- Put a second microchip with channels perpendicular to the first one to add a different set of spatial barcodes that ligate to the first set, creating a 2D grid of pixels containing unique combinations of barcoded info
- Release cDNA and genomic DNA fragments with reverse crosslinking, with cDNAs being captured with streptavidin and gDNA remaining in the supernatant, then sequence everything with NGS
With these approaches, the teams co-mapped the brains of mouse embryos, young mice, and an adult human. They showed that some brain regions are primed during development—the DNA is accessible, but there’s low transcription. In addition, some cell types might retain the ability to differentiate into certain cell types later on because even though transcription had decreased, the DNA was still accessible.
Now, the team says that it should be possible to add yet another layer—proteomics. And the applications could easily go beyond development and neuroscience to figure out disease mechanisms.
Have your multi-layered cake, and eat it too at Nature, March 2023.