Alphafold Made Simple

What is a protein?

Proteins are chains of amino acids that fold into 3D machines. Their shape lets them grab oxygen, copy DNA, digest food, signal cells, or move muscles. Shape comes from chemistry and environment; misfolds can drive disease.

What problem does AlphaFold solve?

Experimental structures are slow and expensive; many proteins had no structure.
AlphaFold predicts 3D shape from sequence and estimates its own confidence.
It points to regions that are solid vs. uncertain, so researchers know what to trust.

Why this matters to humanity

Speeds drug discovery: pick better targets, design antibodies, reduce lab cycles.
Helps rare/orphan proteins: gives first structural clues where no data existed.
Enables greener chemistry: design enzymes for materials, food, and energy.
Democratizes access: anyone can inspect structures without a wet lab.

Key terms (light on jargon)

Sequence: the amino-acid letters (A, C, D...).
Structure: the folded 3D shape that controls function.
pLDDT: per-residue confidence (0–100). Higher = model trusts local shape.
PAE: predicted aligned error between two regions. Low = confident about their relative position; high = uncertain.

Start Here (2–3 minutes)

Pick an example below or enter a UniProt ID.
Rotate the protein and notice the color bands (confidence).
Compare the confidence bars and PAE heatmap.
Skim the notes at right to interpret what high/low confidence means.

Find a Protein

UniProt ID

Examples:

Hemoglobin: carries oxygen; mostly very high confidence.
TP53: genome guardian; flexible tail is low-confidence (likely disordered).
DGAT2: builds triglycerides (fat storage); membrane enzyme.
MAPK1 / ERK2: relays growth signals (starts with “MENFQKVEKIGEGTYGV…”).
FFAR2 / GPR43: senses gut microbe fatty acids; a GPCR.
AT1G58602: plant receptor-like kinase; growth/stress signaling.
Q5VSL9: E. coli membrane/transport-associated protein; prokaryotic example.

Sequence (optional)

Sequence search depends on the AlphaFold API and may return multiple candidates.

3D Structure

Very high (≥ 90) Confident (70–89) Low (50–69) Very low (< 50)

Colors map to pLDDT confidence: blue = trusted shape; orange = uncertain/flexible.

Confidence & PAE

UniProt ID: —

Protein name: —

Average pLDDT: —

Sequence length: —

Model file: —

Confidence breakdown

Predicted Aligned Error (PAE)

Low PAE means the relative positions of two regions are trusted. High PAE means the model is uncertain about their relationship.

Protein Folding Basics

Proteins start as a chain of amino acids.
The chain folds into a 3D shape that enables function.
Folding depends on chemistry, environment, and interactions.

How to Read the Model

Shape: The cartoon shows the folded 3D form that drives function.
Color (pLDDT): Blue = high confidence in local shape; orange = low, likely flexible.
PAE heatmap: Low PAE = two regions are placed confidently; high PAE = their relative position is uncertain.
Disorder: Long low-confidence stretches often mean natural flexibility, not an error.

Try the API (Copy & Paste)

Prediction JSON: https://alphafold.ebi.ac.uk/api/prediction/P69905
UniProt summary: https://alphafold.ebi.ac.uk/api/uniprot/summary/P69905.json
Sequence search: https://alphafold.ebi.ac.uk/api/sequence/summary?id=MVLSPADKTNVKAAW&type=sequence

Data: AlphaFold Protein Structure Database (CC-BY-4.0); please attribute the DB and original papers.

Common Misconceptions

AlphaFold predicts structure; it does not simulate folding physics.
Predictions still need experimental validation in the lab.
Low confidence does not mean wrong sequence; it means uncertainty.
Some proteins only fold when bound to partners.

About the Data

This tool uses the AlphaFold Protein Structure Database API and is read-only. Data is provided under CC-BY-4.0 and requires attribution to AlphaFold DB and the original papers.