Most single-cell sequencing today happens inside a droplet of oil. A machine squirts a cell and a bead into a bubble of water suspended in oil, the bubble seals shut, and inside that sealed space the cell's RNA gets tagged with a barcode. It works, and it scaled the whole field. But a droplet is a dead end for the cell. Nothing gets in, nothing gets out, and the cell cannot survive long inside it. You get one snapshot, then the cell is gone.
A team at Vilnius University wanted a container that behaves less like a sealed bubble and more like a cell's own membrane. According to PubMed, in a paper published in Science on December 18, they describe what they call semipermeable capsules, or SPCs: tiny gel shells that hold a single cell but have walls porous enough to let small molecules pass while keeping large ones like DNA locked inside (doi.org/10.1126/science.ady7227).
Why a leaky wall is the whole point
That selective leakiness is the trick. Enzymes, nutrients, and reaction chemicals can diffuse in and out. The genetic material a cell releases when it is broken open stays put. So the capsule can run a multistep biochemical reaction, the kind that needs several rounds of adding reagents and washing them away, without the cell's molecules escaping and mixing with a neighbor's. Droplets cannot do multistep chemistry easily, because you cannot open and reseal them.
The authors show the capsules handle a range of assays: single-cell genome sequencing, single-cell messenger RNA sequencing, and sorting individual cells by a nucleic acid marker using a standard flow cytometer. That last part matters. Because the capsules are compatible with fluorescence-activated cell sorting, a machine most labs already own, you can fish out the specific transcriptomes you care about instead of sequencing everything and paying for the rest.
The part droplets never managed
Here is the feature that sets the capsules apart. They are biocompatible, so a cell placed inside can keep living. It can divide. The team grew single cells into clonal colonies inside the capsules over long stretches of time. In a droplet, that is impossible; the cell is starved and sealed off from the start. Being able to culture a cell and then sequence it, or watch a clone expand before analyzing it, opens experiments that a one-shot droplet simply forecloses.
The building blocks are ordinary. The capsules are made from biocompatible gel, they work with existing sorters and sequencing chemistry, and they are meant to be, in the authors' framing, easy to use and customizable rather than a bespoke instrument. That is a deliberate contrast with droplet microfluidics, which usually demands specialized chips and pumps.
What the paper does not settle
This is a methods paper, and it should be read as a demonstration of what the platform can do, not proof that it beats the incumbent on every axis. The abstract does not give head-to-head numbers on cost, throughput, or data quality against commercial droplet systems, so those comparisons remain to be worked out by labs that adopt it. Several of the authors are affiliated with Atrandi Biosciences, a company commercializing the approach, which is worth knowing when weighing the framing. And "supports long-term cultivation" is a capability shown in the lab; how well it holds up across many cell types and messier clinical samples is the kind of thing that only becomes clear once other groups put it through its paces.
Still, the direction is interesting. Single-cell biology has spent a decade optimizing a container that kills the cell it measures. A capsule that keeps the cell alive, lets you feed it, and still delivers clean sequencing changes what questions you can even ask. You could track how a single cell's descendants diverge, then read out each one's genome. The oil droplet made the field possible. This looks like an attempt to give it a second container that does not force a trade between measuring a cell and keeping it.
Comments