← Learn

The protein-structure pathway

A good order to try, not a rulebook. These nineteen rungs build on one another: each assumes the one before it, so you can start at the right place for your age and climb. Every rung is a free, no-install tool — some here on Fold Commons, some on our sibling site Spark & Anvil. Nothing is collected.

This is a suggested pathway, informed by curriculum standards and the misconceptions teachers see most — it is not a claim about how much anyone learns. Skip around freely; the arrows just show what tends to make the next rung easier.

Ages 9-14

Start with a cell and its parts, and climb to how a protein's shape does its job and how a prediction can be sure in some places and unsure in others.

  1. BioForge Spark & Anvil

    Cells have parts, and proteins do most of the work.

    Start here Spark & Anvil

    You have to know a protein is a working part of a cell before its shape can mean anything.

    anchored to NGSS LS1.A

  2. Junior explorer Fold Commons

    A protein is a 3D shape, not a line of letters.

    live

    after rung 1: Once a protein is a working part, the first real leap is seeing it as a shape you can turn, not a string.

    anchored to a protein is a shape, not a line

  3. ProteinQuest Fold Commons

    The shape is built from twenty amino acids that differ in character - some shun water, some love it.

    live

    after rung 2: A shape has to be made of something: the twenty amino acids, sorted by whether they hide from water, are the pieces that build it.

    anchored to 20 amino acids differ in character

  4. TraitForge Spark & Anvil

    DNA -> mRNA -> codon -> amino acid; one changed letter can change one amino acid.

    Spark & Anvil

    after rung 3: If amino acids are the pieces, the next question is where they come from - the gene spells them out, one codon at a time.

    anchored to MS-LS3-1 (one letter -> one amino acid)

  5. NotMarbles Fold Commons - with VerbSnap & FlexiFit

    The shape does the job, and shapes move and change when they bind - they are not rigid and not a line.

    live

    after rung 4: Knowing the pieces and where they come from, you can ask what the folded shape actually does - and see that it flexes rather than sitting still.

    anchored to shape does the job; shapes move

  6. VanishPoint Fold Commons - with PencilOrInk & HearTheChain

    A predicted model is sure of its shape in some places and unsure in others - four honest colours.

    live

    after rung 5: Now that shapes matter, you can look at a real prediction and read where the model is confident and where it is only guessing.

    anchored to M03 (confidence, qualitative)

  7. DisorderDuty Fold Commons - with TwoIslands

    "Fuzzy" can mean disordered-and-working, not wrong.

    live

    after rung 6: Once you can read low confidence, the trap is reading it as failure - so next you learn a fuzzy region is often flexible on purpose.

    anchored to M05 (disorder is not error)

  8. MicrobeLab Spark & Anvil

    Proteins recognise each other - antibodies, receptors - the bridge to interfaces.

    Spark & Anvil

    after rung 7: A single flexible shape leads naturally to shapes meeting shapes: how proteins recognise and grip one another.

    anchored to M02 (recognition)

  9. HingeHunt Fold Commons - with BuriedBet

    Residues bury at interfaces, and two domains may or may not "know" where each other sit (a first look at PAE).

    live

    after rung 8: Once proteins meet, you can ask whether the model knows how their parts are arranged relative to each other - the last 9-14 step.

    anchored to M04 (inter-domain placement, PAE)

Ages 15-18

Re-enter with the chemistry of bonds and the hydrophobic effect, then read real confidence numbers, quaternary structure, variants, isoforms, medicine and the blind tests that made prediction trustworthy.

  1. ChemQuest Spark & Anvil - then ReactQuest

    Bonds, polarity, and the hydrophobic effect - the chemistry the older track re-enters on.

    Start here Spark & Anvil

    The 15-18 ladder restarts from the chemistry that makes folding happen at all: why oily parts hide from water.

    anchored to IB B1.2.9 (bonds, polarity, hydrophobic effect)

  2. HelixQuest Spark & Anvil

    Gene sequence and variation -> protein -> evolution.

    Spark & Anvil

    after rung 10: With the chemistry in place, you connect the gene sequence to the protein it encodes and how changes in it drive evolution.

    anchored to HS-LS1-1 (gene -> protein -> evolution)

  3. ShapeForge Spark & Anvil - hands off to the viewer

    Predict local structure; hydrophobics bury; sickle cell is one changed letter. This rung hands off into the Fold Commons viewer.

    Spark & Anvil

    after rung 11: Predicting local structure from sequence is exactly where a real AlphaFold model becomes worth reading - the hand-off point.

    anchored to AP SYI-1.B.2 (predict local structure)

  4. NumberReader Fold Commons - the scored real-numbers read

    Read the real per-residue pLDDT numbers off a structure and say plainly: this is a prediction, not an experiment.

    new

    after rung 12: Having predicted structure, you open a real model and read its actual confidence numbers - the 15-18 numeric companion to the 9-14 colour read.

    anchored to PDB-101 (read the model, part 1)

  5. GripChain Fold Commons

    Several chains assemble, and binding at one site changes another - quaternary structure and allostery.

    live

    after rung 13: Reading one chain's numbers leads to several chains together, and how gripping in one place is felt in another.

    anchored to M13 (quaternary structure, allostery)

  6. DomainCuts Fold Commons - with FoldCompare & WhatsMissing

    Domains and quantitative PAE; RMSD against the real PDB structure; a model omits cofactors and partners.

    live

    after rung 14: With assemblies in mind, you carve domains, read PAE as numbers, check the model against experiment, and notice what it leaves out.

    anchored to PDB-101 (read the model, parts 2-3)

  7. TolerancePaint Fold Commons - with OneLetterLottery

    Most variants are tolerated; AlphaMissense is a research tool, not a diagnosis.

    live

    after rung 15: Knowing what a model shows and omits, you can ask what one changed letter does - and that most changes are harmless.

    anchored to M08 (variants; AlphaMissense research-only)

  8. SpliceSplit Fold Commons

    One gene can make several proteins - isoforms - each folding to its own shape.

    live

    after rung 16: After single-letter changes, the next surprise is that one gene can be spliced into several different folded proteins.

    anchored to M10 (isoforms)

  9. PocketScout Fold Commons - with AntibodyAtlas, PhenoFold & ClinicalAtlas

    Structure -> pocket -> medicine; phenotype -> gene -> structure.

    live

    after rung 17: Once you can read shapes, variants and isoforms, you can hunt the pockets a medicine grips - where the science becomes useful.

    anchored to M02 (structure -> pocket -> medicine)

  10. CASP, explained Fold Commons - with Nobel structures

    Why "prediction" became trustworthy - decades of CASP blind tests, and a Nobel Prize.

    live

    after rung 18: The ladder ends by asking why we trust predictions at all: the blind tests that proved they work.

    anchored to M07 (why prediction is trusted)

How to read this ladder

Each rung assumes the rung before it — that is the arrow. Read them as prerequisites, not grades: rung 6 (reading a model's confidence) is easier once rung 5 (a shape does a job) makes sense, which is easier once rung 3 (the twenty amino acids) does. The two age bands each have their own “Start here” rung: ages 9–14 begin at a cell and its parts; ages 15–18 re-enter at the chemistry of bonds and the hydrophobic effect. The same idea can sit at two rungs at different depths — reading confidence as four honest colours (rung 6, 9–14) and reading the real confidence numbers (rung 13, 15–18) are different rungs, on purpose.

The upstream biology and chemistry rungs live on our sibling site Spark & Anvil, which builds the same free, non-profit way. Fold Commons owns everything from “here is a real protein” onward. Everything AlphaFold shows is a prediction, not an experiment.