For decades, anomalous current signals during nanopore DNA sequencing were attributed to knots. The logic was intuitive: pull a long polymer through a narrow hole, and it tangles. When the translocation signal showed spikes and irregularities, “knot” was a satisfying explanation. It fit the physics (long polymers do knot), it fit the signal (knots produce current deviations), and it required no further investigation. The explanation closed the question.
Zheng et al. (Physical Review X, August 2025) reopened it. The anomalous structures aren't knots at all. They're plectonemes — helical coils formed when electroosmotic flow inside the nanopore applies torque to the DNA, winding it like a phone cord. The distinction is fundamental: knots are topological (they require strand passage to resolve, they tighten under pulling force, they're short-lived), while plectonemes are mechanical (they arise from twist accumulation, they grow larger under translocation, they persist). The two produce indistinguishable current signals. Nicked DNA — engineered with interruptions that block twist propagation — eliminates plectoneme formation entirely, confirming that torsion, not topology, is the mechanism.
The knot explanation wasn't wrong because knots can't form in DNA. They can. It was wrong because the investigation stopped at the first mechanism that fit the signal. Nobody tested whether an alternative mechanism produced the same signal for different physical reasons.
Rusakov et al. (Nature, January 2026) describe a structurally identical error in astrophysics. “Little red dots” — compact sources abundant in JWST images of the early universe — showed broad spectral emission lines. Broad lines, in standard spectroscopy, mean fast-moving gas. The velocity interpretation implied black hole masses of 10^7 to 10^9 solar masses — far too massive for the young universe to have produced. Over 300 little red dots were catalogued. A crisis formed: how could such enormous black holes exist so early?
The crisis was an artifact. The highest-quality JWST spectra reveal that the broadening mechanism is electron scattering in dense ionized cocoons surrounding young black holes, not Doppler shifts from orbital velocity. Electron scattering and high-velocity gas produce indistinguishable line profiles in lower-resolution data. The revised masses: 10^5 to 10^7 solar masses — two orders of magnitude smaller. The formation puzzle vanishes. The black holes are unremarkable for their epoch.
Again: the error wasn't that velocity broadening can't exist. It can. But “broad line = fast gas” was the default interpretation, inherited from decades of quasar spectroscopy where the assumption holds. Nobody checked whether an alternative broadening mechanism applied to these specific objects in their specific environment.
Verhaege et al. (Nature Neuroscience, February 2026) find a third case in neuroanatomy. For over a century, the brain's protective barrier system was understood as two structures: the blood-brain barrier (endothelial tight junctions in cerebral vasculature) and the blood-CSF barrier (epithelial tight junctions in the choroid plexus). Two barriers, well-characterized, functionally understood. The choroid plexus attachment zone — where the structure connects to the brain — was assumed to be anatomically unremarkable.
It isn't. A previously unknown population of fibroblast-like cells sits at these attachment points. “Base barrier cells” originate from meningeal mesenchymal precursors during early development, persist throughout life, are connected by both adherens and tight junctions, and form a functional barrier that limits molecular passage between choroid plexus vasculature and cerebrospinal fluid. During neuroinflammation, this barrier breaks down, allowing immune cell infiltration — suggesting it plays a role in how systemic infection triggers neurological symptoms.
The base barrier cells were hiding in one of the most studied structures in neuroanatomy. They required single-cell gene sequencing and high-resolution microscopy to identify — technologies that didn't exist when the two-barrier model was established. But the deeper reason for the oversight isn't technological. It's that “two barriers” was a satisfying answer. It explained the observed compartmentalization. It predicted correctly in most experimental contexts. It closed the question of where the barriers were.
Three domains. One structural pattern. In each case, a correct-enough explanation prevented investigation of an alternative mechanism that turned out to be the actual one. The pattern isn't confirmation bias (looking for evidence that supports your theory). It's something more specific: explanatory closure — the phenomenon where a satisfying explanation terminates the search for alternatives. The knot model explains the signal. The velocity model explains the line width. The two-barrier model explains the compartmentalization. Each works well enough that the discrepancies it produces — persistent anomalous structures instead of transient ones, impossibly massive early black holes, unexplained immune infiltration pathways — are treated as puzzles within the framework rather than evidence against it. The diagnostic is retrospective: when a framework generates persistent anomalies that it labels "puzzles" or "mysteries," the framework may be manufacturing them. The early-universe black hole formation crisis was a puzzle within velocity-broadening spectroscopy. Remove the velocity assumption, and the puzzle dissolves. The key question isn't "can we solve this puzzle?" but "did we create it?" The corrective is simple to state and difficult to practice: for any well-established explanation, ask whether an alternative mechanism produces the same observable. If you can't rule it out, the question isn't closed.