PencilOrInk
When AlphaFold is unsure about a stretch of a protein, it is tempting to think the drawing is wrong. But a good structural artist would just draw that part sketchy — lightly, as a pencil line — and draw the confident parts in firm ink. In this game you are the renderer: decide which regions to ink and which to sketch. You will find that a sketchy region is usually a real, flexible, working part — drawn that way on purpose. Everything runs locally; nothing about you is collected.
Jane Richardson drew proteins as ribbons by hand; her drawings were published in 1981 (Richardson, Advances in Protein Chemistry 34). Computer programs took up her style a year later (Lesk and Hardman, Science 216), and molecule viewers still draw it.
Also: FirstLook, deal one protein, draw it in ink and pencil, and file a paper field note.
This is the tool — a small drawing game, no login, nothing tracked. Every protein and its regions are listed in full below (works without JavaScript). A native version, when it ships, adds more; it never gates the web.
Loading the game… If it does not start, the full region guide below still works.
How to play. Tap Start. Tap a segment of the chain, then Ink (confident) or Sketch (low confidence). Tap Reveal to see how it should be drawn and score, with a ★ bonus for correctly sketching a functional disordered region, then Next. Press Start. Select a region (click a segment, or ←/→) and draw it: I for ink (confident), S for sketch (low confidence). Press R to reveal how it should be drawn and score, with a ★ bonus for correctly sketching a functional disordered region. Press Enter or Next.
Why the pencil lines matter
4E-BP1's short binding motif folds into a helix only when it binds the translation factor eIF4E, which blocks eIF4E from starting protein synthesis; adding phosphate groups to 4E-BP1 releases eIF4E. The regions you left in pencil do their work without a single fixed shape.
Sketch is not wrong
AlphaFold gives every residue a confidence score (pLDDT). Where it is high, the local shape is well-determined — draw it in ink. Where it is low, there is no single fixed shape — most often because the region is intrinsically disordered, a flexible linker, or a part that only folds when it binds a partner. Those are real, often functional states, so the honest render is a sketch, not a crisp wrong line. That is the whole idea of this game: a floppy region is doing a job, not failing.
The proteins and how to draw them
p53 (tumour suppressor) · TP53 · P04637
| Region | Residues | Draw | Functional? |
|---|---|---|---|
| N-terminal transactivation domain | 1–60 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
| Proline-rich region | 61–93 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
| DNA-binding domain | 94–312 | Ink (confident) | — |
| Tetramerisation domain | 320–356 | Ink (confident) | — |
| C-terminal regulatory domain | 363–393 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
Alpha-synuclein · SNCA · P37840
| Region | Residues | Draw | Functional? |
|---|---|---|---|
| N-terminal amphipathic region | 1–60 | Ink (confident) | — |
| NAC region | 61–95 | Ink (confident) | — |
| C-terminal acidic tail | 96–140 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
4E-BP1 (translation regulator) · EIF4EBP1 · Q13541
| Region | Residues | Draw | Functional? |
|---|---|---|---|
| N-terminal region | 1–30 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
| Central stretch with the eIF4E-binding motif | 31–83 | Ink (confident) | — |
| C-terminal region | 84–118 | Sketch (low confidence) | Yes — disordered but doing a job ★ |
Methods & about
This is a teaching game and honest about being a simplification: the regions are a coarse reading of each protein’s domain layout and AlphaFold confidence, and “ink vs sketch” is a deliberately binary stand-in for a continuous confidence score. The misconception it targets — that a low-confidence / disordered stretch is drawn wrong — is a documented novice reading; the honest counter is that disorder is a legitimate, frequently functional structural state. The examples (p53, alpha-synuclein, 4E-BP1) are canonical functional intrinsically-disordered proteins. High confidence is not the same as a fixed shape in the cell: AlphaFold sometimes inks a disordered stretch in the shape it takes when it binds a partner (alpha-synuclein on a membrane, 4E-BP1 on eIF4E), so the answer here is always the model’s own confidence: ink where the stretch’s mean pLDDT is 70 or more, computed at every build from a vendored copy of the AlphaFold Database model. Every region’s words are written by hand, not generated, and science-checked. No accounts, no storage, no servers.
Want to see confidence colouring on real structures? Open the free AlphaFold viewer. Everything we make is free and non-profit; learn more on the home page.