transcript
0 nt left
0 moves
longest chain — / 20 to win
a 2048 of the central dogma
Swipe to slide the nucleotides. The direction you swipe is the 5′→3′ reading direction — the wall sets the frame, and every three nucleotides in frame translate through the real genetic code.
Tiles compact against the wall you swipe toward. That wall is the 5′ end. The line is then read
outward in non-overlapping triplets, so there is never any ambiguity about how three tiles group —
exactly the reason real translation is unambiguous. Swiping the other way reads the same bases in
reverse: AUG (Met) becomes GUA (Val). Choosing the direction is choosing the codon.
Any three nucleotides in frame translate — they do not have to match. The skill is not finding a merge, it is steering which amino acid you get. A peptide or stop tile sitting mid-line breaks the frame; each segment re-reads from its own leading edge, which is a frameshift you can engineer.
Chains join only within a biochemical class — a conservative substitution. There are two ways, and they are different moves.
Elongation adds a single residue to a chain, and it is directional: the chain must be 5′ and the residue 3′, because a ribosome extends the C-terminus. A residue on the N-terminal side does not join, which is another reason the wall you read from matters.
Ligation fuses two chains of the same length, doubling it — 1→2→4→8 — and scores twice what an elongation to the same size does. It is the fast way up. Either way the tile carries the real concatenated sequence; hover it to read the peptide.
A chain sitting on the board is worth nothing. It banks only when it is terminated, for length² × class rarity — so a 12-mer acidic chain is worth 432 and a 4-mer aliphatic one is 16. There are two ways to terminate.
A stop codon. UAA, UAG and UGA — 3 of the 64 — do not
translate. They drop an inert STOP tile that slides but never merges. Push one into the 3′ end of a
chain and release factor hydrolyses the chain off. Free, but rare: about one stop every twenty codons,
and it has to arrive where you need it.
A release factor. You earn one charge every 12 codons translated. Click any chain of 2 or more residues to spend a charge and bank it on demand. The charges are the scarce thing, so the decision is never whether you may terminate — it is whether this chain is worth a charge, or whether you can grow it further first.
An mRNA is a finite molecule. You get a fixed budget of nucleotides, shown above the board, and when it is spent no more arrive. The endgame is banking everything still on the board before you run out of productive moves. Because every run has the same budget, two scores are comparable.
The ribosome exit tunnel holds roughly 30–40 residues. A chain at 35 cannot elongate further — it has to come off. Hoarding one enormous chain is not a strategy.
Build a chain of the target length shown under the board — 12 on Sprint, 20 on Standard, 28 on Marathon — and bank as much as you can before the transcript is spent. The longest chain that will fit in the tunnel is 35.
The triplet-grouping conflict is a one-dimensional problem — a hex line is just as ambiguous as a square one, so extra neighbours buy nothing. The reading frame is what removes the ambiguity, and it works on any lattice. A 5-wide line leaves a two-base remainder after one codon, which is the frameshift pressure the game is built on.