You can edit a cell perfectly, and it still turns out useless. The gene is fixed exactly as planned, yet the cell behaves unpredictably: it drifts, changes fate, loses safety. It's like fixing one gear in a watch, but the watch still runs wrong because the whole mechanism is out of sync.
Previously, a cell was viewed as a list of independent parts. The authors of this work offer a different perspective: a cell is a ball on a hilly surface. Each valley is a stable state, and hills are barriers between them. After editing, the ball might roll into the wrong valley. The new metric, geometric stability, measures how consistently the cell responds to changes. If all responses point in the same direction, the cell rolls confidently toward the desired valley. If responses diverge, the ball darts around and may get stuck on a slope.
This matters for future medicine, where cells are reprogrammed for therapy. The question shifts from "Did the intervention happen?" to "Is the new state stable?" If a cell is unstable after editing, the therapy may fail or become dangerous. The new method helps foresee where the ball will roll.
The concept of the epigenetic landscape, which underpins this work, was proposed by Conrad Waddington back in 1957. He drew a ball rolling through valleys as a metaphor for cell development.
In the series "Westworld," robot hosts are reprogrammed, but sometimes their behavior becomes unpredictable, like cells that roll into unexpected states after editing.