BuriedBet
For ages 15–18: TwoChains, bet whether a predicted pair of chains is trustworthy, then read the model's own join score and look for an experiment.
AlphaFold usually shows you a protein on its own — but many proteins only work when they lock onto a partner. A whole face of the surface is a hidden interface, buried when the partner arrives. BuriedBet gives you one subunit and asks you to bet which parts get buried; then it loads the real assembly and scores you against the true interface. Everything runs locally; nothing about you is collected.
This is the tool. The interface is computed in your browser from a real experimental assembly (from the RCSB PDB): the reference chain’s residues within 5 Å of any other chain. The proteins are listed below (works without JavaScript).
Loading the game… If it does not start, the full protein guide below still works.
How to play. Press Start. You see one subunit. Select a region (bar: click, or ←/→) and mark it 1 Buried at interface or 2 Stays exposed. The round tells you how many of the 14 stretches are buried; your job is to find them. Press R to load the assembly: you see how many of the buried stretches you found, how many marks were wrong, and how many marking at random would find. Press Enter or Next to continue.
Why the buried face matters
Transthyretin normally circulates as a four-chain assembly; when it comes apart, single chains can misfold into amyloid, and the drug tafamidis binds at the interface between the chains and holds the tetramer together. SOD1 works as a two-chain dimer, and many ALS-linked variants weaken it.
Why the partner matters
A protein’s interface is the patch of surface it buries against a partner. Those residues are usually conserved and often disease-linked, but you cannot see them on the lone subunit — they only make sense once the partner is there. That is the whole point of this game: the monomer is not the whole story. AlphaFold’s 2026 expansion into complexes is exactly about putting the partners back.
The proteins
Answer key — open it after you play
| Protein | Assembly | Reveal (PDB) | Why it needs a partner |
|---|---|---|---|
| Superoxide dismutase 1 (SOD1) P00441 | homodimer | 2C9V | Two identical subunits packed face to face; the buried interface helps hold each subunit stable, and many ALS-linked mutations sit near it. |
| Triosephosphate isomerase (TIM) P60174 | homodimer | 1HTI | An obligate homodimer — each subunit has its own complete active site, but the tightly packed interface between them is crucial for stability and full activity. The monomer alone is not the working enzyme. |
| Transthyretin (TTR) P02766 | tetramer | 1TTA | Works as a tetramer (a dimer of dimers) carrying thyroid hormone; amyloid disease begins when the assembly comes apart. The reveal shows the strong dimer interface. |
Methods & about
The lone subunit is AlphaFold’s monomer model (from the AlphaFold Database); the assembly is a real experimental structure fetched from the RCSB PDB. The interface is computed in your browser as the reference chain’s residues within 5 Å (heavy atom) of any other chain, using a shared mmCIF parser — the same kind of contact calculation our other structure tools use. It is a geometric contact interface, a teaching approximation, not a curated interface annotation; for tetramers (transthyretin) the deposited coordinates contain the strong dimer — the full tetramer is completed by crystal symmetry — so we reveal and score that dimer interface. Every protein note is written by hand, not generated, and science-checked. No accounts, no storage, no servers of ours.
Want to explore assemblies yourself? Open the free AlphaFold viewer or FoldCompare. Everything we make is free and non-profit; learn more on the home page.