Click or drag files to upload (.pdb, .ent)
Aggrescan3D (A3D) predicts which regions of a protein are prone to aggregation based on its three-dimensional structure. While sequence-based aggregation predictors treat each residue independently, A3D accounts for the fact that residues far apart in sequence can cluster together on the protein surface, forming aggregation-prone patches that only become apparent in 3D.
The method builds on the AGGRESCAN amino acid aggregation propensity scale, derived from experimental measurements of amyloid-beta peptide mutants expressed in E. coli. Each residue's intrinsic aggregation tendency is then corrected for two structural factors: how exposed it is to solvent, and what other residues lie nearby in space. The result is a per-residue A3D score that distinguishes surface-exposed aggregation hotspots from buried hydrophobic cores that are structurally necessary but not aggregation-relevant.
A3D calculates a structurally corrected aggregation propensity for each residue in three steps:
Intrinsic propensity assignment: Each amino acid receives a value from the experimentally derived AGGRESCAN a3v scale. Hydrophobic residues like phenylalanine (+1.27), isoleucine (+0.86), and valine (+0.75) score positive (aggregation-prone), while charged residues like lysine (-1.48), arginine (-1.01), and aspartate (-0.72) score negative (solubility-promoting).
Solvent accessibility weighting: Residues buried inside the protein contribute little to surface-driven aggregation. A3D computes the relative solvent accessible surface area (RSA) for each residue and applies an exponential weighting function. Residues below 10% RSA are effectively excluded; fully exposed residues receive full weight.
Spatial aggregation: For each residue, A3D considers all neighboring residues within a sphere centered on its alpha-carbon. The default radius is 10 Å, which captures aggregation patches formed by spatially clustered residues. Contributions from neighbors decay with distance. A smaller 5 Å radius isolates individual residue contributions without patch-level effects.
The final A3D score for each position reflects whether a residue sits in a context that promotes aggregation (positive score) or solubility (negative score).
ProteinIQ runs Aggrescan3D directly in the browser against PDB structures, with no software installation or command-line setup required.
| Input | Description |
|---|---|
Protein Structure | One or more PDB files, or fetch structures by PDB ID (e.g., 1CRN). Supports multi-chain and multi-model structures. |
| Setting | Description |
|---|---|
Distance threshold | Radius of the spatial sphere around each residue. 10 Å (default) detects aggregation patches from clustered residues. 5 Å reports individual residue contributions. |
Results are presented across three tabs:
| Output column | Description |
|---|---|
Structure | PDB ID or filename |
Chain | Chain identifier |
Residue # | Residue sequence number |
Residue | One-letter amino acid code |
A3D Score | Structurally corrected aggregation propensity |
Accessibility | Relative solvent accessible surface area (%) |
Classification | Aggregation-prone, Neutral, or Solubility-promoting |
| A3D Score | Classification | Meaning |
|---|---|---|
| > 0 | Aggregation-prone | Residue sits in a surface-exposed context that favors aggregation. Higher values indicate stronger aggregation tendency. |
| ~ 0 | Neutral | Residue has minimal impact on protein aggregation from its structural context. |
| < 0 | Solubility-promoting | Residue is in a context that actively disfavors aggregation, typically due to nearby charged or polar residues. |
The average A3D score across all residues serves as a global indicator of aggregation propensity, enabling comparison between different proteins or mutant variants. More negative averages indicate higher overall solubility.
Clusters of positive-scoring residues on the protein surface are the most actionable finding. These aggregation-prone patches are candidates for solubility engineering through targeted mutation to charged amino acids (Arg, Asp, Glu, Lys).
