Playable

De Novo

Two things have to happen at a locus before a gene exists: it has to be transcribed, and it has to contain an open reading frame. Neither is stable. Both decay while you attend to the other. And protein — the only thing selection can see, and the only score here — is made when both are up.

Q W alternate to drive RNA polymerase. O P alternate to extend the frame. Tap the on-screen keys if you prefer.

0peptides translated

non-genic DNAintergenic sequence, doing nothing

TranscriptionQ · W0%

falls when a nucleosome closes over the promoter

Reading frameO · P0%

falls when a premature stop codon drifts in

transcription

frame

Press Start.

De Novo

Alternate Q/W to transcribe. Alternate O/P to lengthen the reading frame. The marked key is the one that will count; pressing the same key twice stalls rather than kills. Protein accrues only while both meters are up, and the ladder on the right is denominated in peptides.

Difficulty changes how fast the two meters decay and how long you can produce nothing before selection deletes the locus. There is no penalty for Assist and nothing is withheld on it.

Rules, in full

Input

Each pair must alternate. Pressing Q twice in a row does not advance the polymerase — it stalls for 0.16 s, which costs about one decay event's worth of progress. It is never fatal on its own. Holding a key down does nothing; auto-repeat is ignored deliberately.

Decay

Both meters fall in discrete visible events, not as a silent drain: a nucleosome closes over the promoter, or a premature stop codon appears in the frame. If a meter went down, something happened on screen.

Score

One number: peptides. The rate is 26 × transcription × frame per second, so 100% of one and 0% of the other pays exactly nothing. Every rung on the ladder is a peptide threshold, so progress and score are the same unit.

Losing

Produce nothing for the grace window and purifying selection removes the locus. The countdown bar appears the instant the timer starts — you always see it coming, and it always resets the moment you produce again.

Escalation

Each rung shortens the decay periods. Holding both meters costs about 2.2 presses per second at the bottom of the ladder and about 7.1 at the top. The rhythm that got you started will not get you fixed.

Which meter to feed

Transcription drains faster early; the frame drains faster late. Feeding them 50/50 works, but tracking the skew finishes measurably faster — about 42 s versus 51 s at the same press budget. It does not change whether you survive.

What the game gets right, and what it distorts

It is a game, so it lies. Being specific about which lies it tells is more useful than pretending it doesn't.

Mechanic	The biology it stands for	The distortion
Two independent meters	Transcription and ORF length really are separable properties, acquired by different mutations	Fair. This is the game's one genuinely load-bearing claim.
Score = T × F	A transcript with no ORF makes no protein; an ORF that is never transcribed makes no protein	Fair, though real translation is far more switch-like than multiplicative.
Both meters decay	Mutation degrades unconstrained sequence; promoters and ORFs both erode without selection	Real erosion runs over millions of years, not seconds, and does not tick at a fixed rate.
Order of acquisition names the scenario	ORF-first versus expression-first is a genuine, argued distinction in the literature	Fair as a teaching device. In real data the ordering is inferred from outgroups, not observed.
You control the locus	—	The big lie. Nothing steers this process. Every mutation is undirected, and the filter runs afterwards. The game gives you agency the ancestor did not have.
Climbing the phylostratum ladder	Fixation, then persistence long enough to be detectable at deeper nodes	Backwards. A gene does not climb strata over its lifetime — the stratum is where the observer can still detect it today. Here it stands in for elapsed time.
Survival ≈ 6.5 presses/s	Purifying selection removing what makes nothing	Real proto-genes almost all die. The game's win rate is wildly optimistic.

The last row is the one worth sitting with. If the game were honest about the odds it would delete your locus, without warning, about 999 times out of 1000 — which is the actual shape of step 5 and a terrible game.