The standard way to test whether a genebanked seed is still alive is to plant it and see. This destroys the seed. For endangered species with collections of a few hundred seeds — each one irreplaceable — viability testing creates an agonizing tradeoff: test and lose material, or don't test and risk the entire collection dying silently in the freezer.
Walters and colleagues measured RNA integrity in seeds from over 100 endangered U.S. species stored at -18°C for approximately 28 years. RNA integrity number (RIN) values decline before viability loss becomes detectable through germination — the molecular signal precedes the functional signal. And the RNA method requires as few as 22 seeds, where germination assays would consume hundreds.
The RNA isn't maintaining viability — it's degrading alongside it, as a parallel process driven by the same oxidative chemistry. But the degradation of RNA is measurable at a stage when the seed can still germinate. The molecular clock runs ahead of the functional clock. By the time a germination test detects failure, the seeds that remain are also failing — you've consumed the surviving fraction to measure the dying one.
This transforms seed banking from a faith-based operation to one with early-warning instrumentation. A genebank manager can sample RNA from a small fraction of the collection, detect declining integrity, and intervene (move to colder storage, attempt regeneration) before viability drops below the point of recovery.
The through-claim: when measuring the system destroys the system, the measurement method is part of the problem. A non-destructive proxy that reads degradation before functional failure converts an either-or tradeoff (test or preserve) into a monitoring protocol that does both. The key was not better germination tests — it was finding a signal that runs ahead of the failure it predicts.