Amplification and concentration attack the same visibility problem from opposite ends. The older route, as the Stanford announcement describes it, pulls protein out of many cells and concentrates the one you care about until there is enough of it to see [4]. The currency you spend doing that is cells, and cells are the scarce reagent when the plate has hundreds of wells and each well needs its own drug or its own target [3][5]. Banik's group left the detector alone and rebuilt the thing being detected: native cellular machinery is programmed to emit a signal, so a small drop in protein abundance shows up as a large change in output [8]. Banik calls it one of the first methods to use signal amplification for small abundance changes [6].
The engineering asymmetry Banik points to is the interesting part. Making a signal vanish is easy, and making one appear requires more steps to go right [11]. That is also the specificity problem in miniature. Anything that perturbs the coupling between the target protein and the reporter output gets amplified with the same gain as the biology you wanted, so an amplifying screen lives or dies on its hit confirmation rate rather than on its raw sensitivity.
The numbers are missing from this account. The phys.org write-up carries no fold-amplification figure, no limit of detection, no false-positive rate, and no count of proteins or drugs actually run, and its description of the mechanism stops mid-sentence at programming native cellular machinery [12]. The comparison on offer is internal: the same large signal at small cell numbers as at large ones [7]. That is a scale-invariance claim, measured against itself rather than against a named existing assay [13]. Useful, but it stands as a scale claim rather than a sensitivity benchmark.
Worth separating the motivation from the result. Small decreases in protein level can change what a cell does, and degrading a malfunctioning protein with a drug is a real therapeutic strategy in cancer [10]; none of that is what was demonstrated here. What was demonstrated is a readout, in cells, built and run inside a facility where the high-throughput steps are automated and a screening specialist, co-author David Solow-Cordero, helped adapt the chemistry to that format [9]. A group without that automation gets the sensitivity argument but not the throughput one.
For a degrader program deciding whether to build this, the number to look for in the Cell paper is the reporter's dose-response against a degrader with a known potency, next to the fraction of primary hits that survive an orthogonal protein measurement.
