Amino Acid Hydrophobicity Scales
An overview of major hydrophobicity scales including Kyte-Doolittle, Eisenberg, and Wimley-White, and their applications in membrane protein prediction.
Amino Acid Hydrophobicity Scales
Hydrophobicity scales quantify the preference of amino acids for aqueous versus nonpolar environments. These numerical assignments are essential for predicting protein structure, identifying membrane-spanning regions, and designing peptides with specific solubility properties.
Major Hydrophobicity Scales
Kyte-Doolittle Scale
The Kyte-Doolittle scale (1982) assigns values from -4.5 (most hydrophilic, arginine) to +4.5 (most hydrophobic, isoleucine). Values are derived from experimental free energy of transfer from water to vapor. This scale remains widely used for hydropathy plots that predict transmembrane domains.
Mnemonic: “I am Very hydrophobic” (Ile, Val at the top); “R is Really water-loving” (Arg at the bottom).
Eisenberg Scale
The Eisenberg consensus scale normalizes values from multiple experimental measurements, including octanol-water partitioning and crystal structure analysis. Values range from -1.6 (glutamate) to +0.9 (isoleucine). This scale is particularly useful for calculating hydrophobic moments that predict amphipathic helices.
Wimley-White Scale
The Wimley-White scale (1996) measures whole-residue free energies of transfer from water to n-octanol or interface environments. This scale better reflects membrane protein behavior because it accounts for the interfacial environment rather than bulk hydrocarbon phases.
Applications in Membrane Protein Prediction
Hydropathy analysis uses a sliding window (typically 19-21 residues for alpha-helices) to identify stretches of high hydrophobicity. Three consecutive hydrophobic residues with Kyte-Doolittle values above +1.6 typically indicate a transmembrane segment.
Choosing the Right Scale
- Kyte-Doolittle: Quick hydropathy plots, general solubility predictions
- Eisenberg: Amphipathic helix detection, hydrophobic moment analysis
- Wimley-White: Membrane protein energetics, lipid interface behavior
Learning Tip
No single scale is universally superior. Cross-validate predictions using at least two scales. When a region scores high on multiple scales, confidence in the prediction increases substantially.