Key Takeaways
- FIND, from Wesley Wong's lab at Boston Children's, Harvard Medical School and the Wyss Institute, screens molecule libraries using DNA nanoswitches that change shape when two prote
- It demonstrated nanobodies that stabilize the ACE2–SARS-CoV-2 spike interaction and PPARγ agonists from a DNA-barcoded library described as 400 million compounds.
- FIND is a laboratory discovery platform, not a medicine — its hits still require years of development and clinical testing.
Most drugs block something. These would glue things together
Modern drug discovery is very good at finding molecules that stop a protein from working. It is far worse at finding molecules that make two proteins interact — the so-called "molecular glues" that can, for example, drag a disease protein to the cell's disposal machinery. Those interaction-changing molecules are among the most valuable in medicine, and among the hardest to screen for, because conventional screens mostly test whether a molecule binds one protein, not whether it changes a relationship between two.
A team led by Wesley Wong's lab at Boston Children's Hospital, Harvard Medical School and the Wyss Institute has built a way to test for exactly that. Their platform, FIND (Functional Interaction-based Nanoswitch Discovery), was published in Science on November 5.
A switch, a barcode, and a very big library
The trick is a DNA nanoswitch: a strand of DNA engineered to sit in one of two shapes — looped or straight — depending on whether two proteins attached to it are interacting. Each candidate molecule in a screening library is attached to its own nanoswitch and tagged with a DNA barcode that identifies it. Run the mixture, separate the switches by shape, then sequence the barcodes of the ones that flipped — and you have a list of molecules that induced or strengthened a protein–protein interaction. Because every experiment runs in parallel on individual DNA molecules in a single tube, the throughput is enormous: in one demonstration, the team screened a DNA-barcoded small-molecule library described by the institute as holding 400 million compounds.
The platform proved itself on two targets. It found nanobodies that stabilize the interaction between the human ACE2 receptor and the spike protein of SARS-CoV-2 — glue, in other words, for a virus–host handshake. And it found agonists of PPARγ, a diabetes-relevant target, that work by promoting the protein's interaction with a partner peptide.
Why it matters — and what it is not
Interaction-inducing drugs are a frontier because they open targets that blockers cannot reach: proteins with no pocket to block can still be glued, stabilized or escorted to destruction. A screen that tests for interaction effects directly, at this scale, turns a class of discoveries that used to be serendipity into something closer to a search. The technology has already been licensed to a startup, Merge Bio Therapeutics, co-founded by members of the team.
The honest limits: FIND is a laboratory discovery platform, demonstrated in vitro and in cellular assays. Nothing it has found is a medicine yet — every hit still faces years of chemistry, animal testing and clinical trials, and most will fail along the way. What has changed is the speed of the first, hardest step: knowing which molecules are worth that journey.
The Twikup Take
Drug discovery's bottleneck has never been imagination; it is testing. FIND tests the one property — "does this molecule change how two proteins behave together?" — that the industry's standard screens largely cannot, and it tests it millions of times at once. The medicines, if they come, are years away. The capability is here now.
Image: Illustrative image.
