Fold Commons

PhenoFold

What does a genetic disease look like at the level of the protein? Pick a well-established monogenic phenotype below; this tool fetches its causal protein's AlphaFold structure in your browser and shows the structural region involved — the genotype↔phenotype↔structure link, made concrete. Every entry is sourced (OMIM/MONDO, UniProt, and a peer-reviewed paper). Everything runs locally; nothing is installed and nothing about you is collected.

This is the tool. It shows no phenomic images. It maps a curated phenotype to its causal protein's predicted structure, so it is not a genotype-to-phenotype predictor and not a diagnostic. A native version, when it ships, adds curated phenomic datasets (BBBC, JUMP-CP), reading-tier copy, and offline use; it never gates the web.

Coverage

≈5% a curated starter set of well-established monogenic examples; polygenic and complex phenotypes are out of scope.

Known gaps (5)
  • Polygenic and complex phenotypes (height, most common disease risk) are entirely out of scope
  • Most of the ~7000 known rare monogenic disorders are not yet curated here
  • Non-coding / regulatory variants with no direct protein-structure consequence
  • Copy-number, structural, and mitochondrial-genome variants
  • Phenomic imaging data (e.g. cell-painting) is deliberately not hosted — this edition maps phenotype to protein structure, not to images

Pick a phenotype

Causal protein structure

AlphaFold prediction of the causal protein key residue/region highlighted where confidently known · a predicted model, not experimental

Phenotype → gene → protein

Achondroplasia

FGFR3 · P22607

Short-limbed short stature with characteristic facial features (most common form of dwarfism)

  • System Skeletal
  • Inheritance autosomal dominant
  • Region Gly380Arg in the transmembrane domain (residue 380)

Almost all achondroplasia is caused by a single recurrent substitution, Gly380Arg, in the transmembrane segment of fibroblast growth factor receptor 3. The change is generally understood to stabilise receptor dimers and produce constitutive (ligand-independent) signalling that suppresses growth-plate chondrocyte proliferation — a gain-of-function acting on receptor activation, not a loss of the receptor's fold.

Cystic fibrosis

CFTR · P13569

Thick airway mucus, chronic lung infection, and pancreatic insufficiency

  • System Respiratory
  • Inheritance autosomal recessive
  • Region Phe508del in nucleotide-binding domain 1 (NBD1) (residue 508)

The most common CF allele deletes a single phenylalanine (Phe508) from the first nucleotide-binding domain of the CFTR chloride channel. The residue sits at an NBD1 surface that contacts the membrane-spanning domains, and its loss is generally described as destabilising domain folding and assembly so that most protein is degraded before reaching the cell surface — a folding/trafficking defect rather than a change to the isolated NBD1 fold.

Duchenne muscular dystrophy

DMD · P11532

Progressive childhood-onset muscle weakness and wasting

  • System Muscular
  • Inheritance X-linked recessive
  • Region Whole protein — most alleles are frameshifting deletions abolishing dystrophin

Duchenne muscular dystrophy usually results from out-of-frame deletions in the very large dystrophin gene that abolish production of the dystrophin protein, which normally links the muscle-cell cytoskeleton to the membrane. Because the common mechanism is loss of the whole protein rather than a single substitution, no single key residue is marked. Dystrophin is a long rod of spectrin-like repeats; its predicted model is low-confidence in many regions — interpret pLDDT before drawing conclusions about any part of the fold.

Familial hypercholesterolaemia

LDLR · P01130

Very high LDL cholesterol from birth; premature coronary artery disease

  • System Cardiovascular
  • Inheritance autosomal dominant
  • Region Ligand-binding and EGF-precursor domains (many variants across the receptor)

Familial hypercholesterolaemia is most often caused by variants in the LDL receptor, which clears LDL cholesterol from the blood. Hundreds of variants are known across the receptor's cysteine-rich ligand-binding and EGF-precursor domains; they are generally described as impairing folding, ligand binding, or trafficking so that fewer functional receptors reach the cell surface. No single residue dominates, so this entry marks no single key residue.

Huntington disease

HTT · P42858

Progressive chorea, cognitive decline, and psychiatric change in mid-adult life

  • System Nervous system
  • Inheritance autosomal dominant
  • Region N-terminal polyglutamine (polyQ) tract (expansion beyond ~36 CAG repeats)

Huntington disease is caused by expansion of a CAG-repeat that encodes a polyglutamine tract near the N-terminus of the huntingtin protein. Beyond roughly 36 repeats the expanded tract is generally understood to promote misfolding and aggregation of N-terminal fragments; huntingtin is very large and largely flexible, so this is a toxic gain-of-function of an expanded low-complexity region rather than disruption of a well-defined globular fold. The predicted model is low-confidence across much of the chain — read pLDDT before over-interpreting any region.

Marfan syndrome

FBN1 · P35555

Tall stature, long limbs, lens dislocation, and aortic root dilatation

  • System Connective tissue
  • Inheritance autosomal dominant
  • Region Calcium-binding EGF-like domains (many cysteine-affecting variants)

Marfan syndrome is caused by variants in fibrillin-1, a large extracellular matrix protein built from many tandem calcium-binding EGF-like domains. Many pathogenic variants alter conserved cysteines that form the disulfide bonds holding each small domain together; these are generally understood to misfold individual domains and disrupt microfibril assembly. No single residue dominates across patients, so this entry marks no single key residue. The predicted model of this long, repetitive, largely extracellular protein is low-confidence in many regions.

Phenylketonuria (PKU)

PAH · P00439

Elevated blood phenylalanine; untreated, causes intellectual disability (managed by low-Phe diet)

  • System Metabolic
  • Inheritance autosomal recessive
  • Region Catalytic and regulatory domains (many disease alleles; most reduce enzyme stability/activity)

PKU results from reduced activity of phenylalanine hydroxylase, the enzyme that converts phenylalanine to tyrosine. Hundreds of different variants are known across the enzyme; many are generally described as destabilising the folded protein or its assembly so that functional enzyme is lost. Because no single residue dominates, this entry marks no single key residue — the mechanism is broadly loss of a stable, active enzyme.

Sickle cell disease

HBB · P68871

Sickle-shaped red blood cells; chronic haemolytic anaemia and vaso-occlusive crises

  • System Blood
  • Inheritance autosomal recessive
  • Region β6 (Glu→Val; residue 7 in UniProt numbering, which includes the initiator Met) (residue 7)

The classic β-globin substitution Glu6Val (position 6 in the mature chain; residue 7 in UniProt canonical numbering, which counts the initiator methionine) places a hydrophobic valine on the molecule's surface. This surface patch lets deoxygenated haemoglobin S polymerise into fibres, which is generally understood to deform the red cell — the folded monomer is little changed, so the effect is on assembly rather than on the single-chain fold.

Tay-Sachs disease

HEXA · P06865

Progressive neurodegeneration in infancy with a cherry-red macular spot

  • System Nervous system
  • Inheritance autosomal recessive
  • Region α-subunit of β-hexosaminidase A (many alleles reduce enzyme activity/stability)

Tay-Sachs disease is caused by deficiency of the α-subunit of β-hexosaminidase A, so the GM2 ganglioside it normally degrades accumulates in neurons. Multiple alleles reduce enzyme activity or stability by different means; because no single residue dominates across patients, this entry marks no single key residue. The mechanism is broadly loss of a stable, active enzyme.

Methods & limits

About this tool

Structures come from the AlphaFold Protein Structure Database (EMBL-EBI / Google DeepMind). They are fetched directly by your browser and rendered with Mol* via PDBe Mol*, self-hosted here — no third-party CDN, no tracking, no server. Phenotype → gene → protein mappings are curated from OMIM, UniProt, and the primary literature; please cite those and the AlphaFold DB when you use this tool. It is part of the free, non-profit Fold Commons project.