The Kyte-Doolittle scale assigns a hydrophobicity score to each amino acid. Using a sliding window, this plot reveals hydrophobic regions (membrane anchors, signal peptides) and hydrophilic surface-exposed regions.
Scores range from +4.5 (Ile, most hydrophobic) to −4.5 (Arg, most hydrophilic). Kyte and Doolittle combined water-to-vapour transfer free energies of the side chains (Wolfenden et al. 1981) with how often each residue is buried in proteins of known structure, adjusting the values where the two sources disagreed.
The classic application of this plot. Kyte and Doolittle (1982) found that, with a window of 19 residues, segments averaging above 1.6 are likely to be transmembrane α-helices. The window matches the ~20 residues an α-helix needs to cross the 30 Å hydrophobic core of a lipid bilayer. Select window 19 and the 1.6 threshold appears as a dashed green line; a genuine transmembrane segment stays above it over several consecutive positions. GPCRs, for example, show seven such peaks.
An amphipathic helix has one hydrophobic face and one hydrophilic face. On the plot, these appear as moderate oscillations rather than a sustained peak. Melittin is the textbook example — its N-terminal region lies parallel to cell membranes, inserting one hydrophobic face into the bilayer. Most antimicrobial peptides (cecropins, magainins, LL-37) are amphipathic helices; this geometry lets them disrupt bacterial membranes while sparing host cells.
Signal peptides are N-terminal sequences (15–30 residues) that direct newly made proteins into the secretory pathway. Their hydrophobic core of 7–15 residues produces a sharp peak at the very start of the plot, followed by a drop where the mature protein begins; signal peptidase later cleaves the segment off. Try the Signal peptide example above. Note that a signal peptide and an N-terminal transmembrane helix look alike on a hydropathy plot; telling them apart needs a dedicated predictor such as SignalP or Phobius.
In-depth guide: how the Kyte–Doolittle scale was built, how to choose the window size, when the 1.6 threshold applies, what GRAVY tells you, and why amphipathic helices and short peptides are easy to misread — with worked examples.
Read the guide →