Self-blinking DNA dyes push chromatin imaging into the wax blocks hospitals already keep

Blinking is the whole trick, and it is worth being clear about why. Single-molecule localisation only works if most labels are dark at any given instant, so the microscope can fit a position to one isolated flicker instead of a smear; the usual route to that state is harsh chemistry and laser power that living cells do not survive, which is why almost every published image of DNA folding comes from cells that were already dead ^[12]. The HoTs probes enter living cells unaided, bind DNA and flicker on and off by themselves ^[1], and thousands of such frames are reconstructed into an image the authors describe as ten times sharper than a conventional one ^[14]. Pia Cosma, the senior author, frames the payoff as one dye doing two jobs: motion in a live cell, geometry in a preserved one ^[13]. Hongyan Sun, co-senior author, says tuning the blinking behaviour was the hard part ^[15].

The two headline numbers come from two microscopes and two states of matter. STORM gave 20 nanometres inside living cells, against roughly 200 for an ordinary lens ^[4][5]; MINFLUX on chemically preserved laboratory cells pinned single dye molecules to within 3 nanometres ^[6]. Divide one by the other and the live-cell figure is about 6.7 times coarser than the fixed-cell one ^[19]. And 3 nanometres sits at 1.5 times the roughly 2-nanometre width of the double helix ^[7][20]. That is genuinely close to the object, but localisation precision is the uncertainty on where one dye sits, not the distance at which two neighbouring strands can be told apart. The thing it does not tell you is how densely the chromatin was labelled: an unlabelled stretch of DNA is invisible at any precision, and sparse labelling makes packing look looser than it is.

Which matters for the clinical panel, because looseness is exactly what was measured there. In wax-preserved bowel slices from three cancer patients, tumour DNA looked more spread out than the healthy tissue immediately alongside it ^[8][9]. The design choice is a good one: comparing tumour against adjacent normal from the same patient holds fixation, storage and staining constant in a way a comparison across patients would not. The direction also matches earlier work, in which DNA unpacks steadily as cancer takes hold ^[10]. That agreement is reassuring and is also the reason for restraint, since looser chromatin travelling with tumour tissue does not show that folding state drives progression.

The incumbent is not a resolution number. Pathologists read these blocks by eye with a stain more than a century old ^[11], and folding is a plausible thing to add because how tightly the two metres of DNA in a cell is packed governs which genes are readable ^[18]. A new readout competes on throughput and reproducibility across labs, which is not what a resolution figure measures.

My read: the reusable asset here is the chemistry. A probe that crosses into a live cell on its own and blinks without being forced to travels to other labs and other questions, and the three-patient observation is best read as evidence that the chemistry survives wax.