Hydropathy plot

How to make a hydropathy plot?

Paste a protein sequence, FASTA records, or a PDB file, choose a window size, and the Hydropathy plot averages Kyte-Doolittle values over every window of that many residues. Each average is plotted at the window's center residue, so hydrophobic stretches such as membrane-spanning helices rise above zero and surface loops fall below it.

A 10-residue peptide with a 5-residue window gives six points, one for each full window:

>peptide
MKTAYIAKQR

The first value is the mean of M, K, T, A, and Y: (1.9 - 3.9 - 0.7 + 1.8 - 1.3) / 5 = -0.44. Positions 1, 2, 9, and 10 have no full window, so they have no value. ExPASy ProtScale reports the same values.

Finding a transmembrane helix

The bundled example is human glycophorin A (UniProt P02724), a red blood cell protein with one membrane-spanning helix. With a 19-residue window, the plot peaks at 2.67 at position 101 and stays above 1.6 from position 99 to 106:

That peak lies inside the annotated transmembrane helix (residues 92 to 114). A second, shorter rise above 1.6 at positions 10 to 14 is the signal peptide, which is also hydrophobic but is cleaved before the protein reaches the membrane.

Input

Header lines are kept whole as sequence names. A name on the line above an unheaded sequence (such as Human insulin) is used as its name rather than read as residues. Spaces, position numbers, and stop markers (*) are removed; the page notes when stop markers or alignment gaps were dropped, since positions then count residues only. PDB files contribute one sequence per chain.

Sequences shorter than the window are skipped with a note. If nothing is long enough, the run stops and says which sequence was too short. A sequence that looks like DNA still runs but is flagged, because A, C, G, and T are also amino acid letters; translate it with DNA to Protein first.

Settings

Short windows (5 to 9) keep local detail such as surface loops and turns. Use 19 to look for transmembrane helices, the window Kyte and Doolittle paired with their 1.6 cutoff.

Results

The chart can be exported as an image from the camera button.

Rare residue codes follow ProtScale: B scores as the mean of D and N, Z as the mean of E and Q, and X as the mean of all 20 values (-0.49). Selenocysteine (U), pyrrolysine (O), and J have no Kyte-Doolittle value, so windows containing them average over their other residues, and the page lists them.

What is the Kyte-Doolittle hydropathy index?

Kyte and Doolittle (1982) gave each amino acid a hydropathy index from two kinds of evidence: the free energy of moving its side chain from water to vapor, and how often it is buried inside folded proteins. The index runs from 4.5 for isoleucine, the most hydrophobic residue, to -4.5 for arginine, the most hydrophilic. These are the values the Hydropathy plot averages.

Values are from Table 2 of Kyte and Doolittle (1982) and match the Kyte-Doolittle scale in ExPASy ProtScale. Averaging them over a window turns single residues into regional character: hydrophobic stretches tend to be buried in the protein core or a membrane, and hydrophilic stretches sit on the surface in contact with water.

How other hydrophobicity scales differ

Every scale agrees that isoleucine, leucine, valine, and phenylalanine are hydrophobic and that arginine, lysine, aspartate, and glutamate are not. They disagree on the residues in between, because each comes from a different measurement. Tryptophan is the clearest case: Kyte and Doolittle rank it 11th of 20, the Eisenberg consensus scale 5th, and the Fauchere-Pliska octanol scale 1st, the most hydrophobic residue of all.

The scales also span different ranges, so a cutoff calibrated on one is meaningless on another. The Hydrophobicity plot draws the same kind of profile on 24 other scales, including Eisenberg, Fauchere-Pliska, and the Hopp-Woods hydrophilicity scale used for antigenic sites.

How to read a hydropathy plot

Start with the window. Each point is the mean of that many residues, so the window decides what size of feature can stand out. A 5- to 9-residue window shows turns, loops, and short buried strands, but a single isoleucine or leucine can push it up on its own. A 19-residue window averages away that noise and keeps only stretches long enough to cross a membrane.

• Above 1.6 with a 19-residue window: a likely transmembrane helix.
• Above 1.6 near the start: often a signal peptide or signal anchor rather than a membrane helix.
• Below 0: surface-exposed loops and termini, where hydrophilic residues contact water.
• Alternating small peaks and dips: can indicate an amphipathic helix, which has one hydrophobic face and lies along the membrane surface without crossing it.

Hydropathy alone cannot separate these cases or give membrane topology. Treat peaks as candidates and compare them with a structure prediction or a dedicated topology predictor.

Where the 1.6 threshold and 19-residue window come from

Both numbers are from the original paper: Kyte, J. and Doolittle, R. F. (1982), "A simple method for displaying the hydropathic character of a protein", Journal of Molecular Biology 157(1):105-132 (PubMed 7108955). They proposed that a 19-residue segment averaging above 1.6 likely spans a membrane, since that cutoff separated the known membrane-spanning segments in their set from the most hydrophobic buried segments of soluble proteins.

The window length has a physical reason. An α-helix rises 1.5 Å per residue, so 19 residues make a helix about 28 Å long, close to the thickness of a lipid bilayer's hydrocarbon core. The threshold belongs to that pairing: a shorter window averages fewer residues, so short hydrophobic patches in soluble proteins can cross 1.6, and on any other scale the number has no meaning.

Hydropathy plot and GRAVY

The GRAVY calculator reports the grand average of hydropathy: the Kyte-Doolittle index averaged over every residue of the protein. It is the hydropathy plot collapsed into one number, and the collapse hides exactly what the plot shows. Glycophorin A has a GRAVY of -0.080, mildly hydrophilic, even though its 19-residue profile peaks at 2.67 over the membrane helix. One helix in a 150-residue protein barely moves the global mean.

The average under the chart is not GRAVY either. It is the mean of the plotted window values, and residues near the ends fall into fewer windows than residues in the middle, so they count less. For glycophorin A it is -0.034 with a 9-residue window and -0.102 with a 19-residue window. Use GRAVY when you need the whole-protein value to compare proteins.

Which protein property tool should I use?

FAQ

What window size should I use for a hydropathy plot?

Use 19 to look for transmembrane helices and 7 to 11 for surface regions and loops. The default of 9 matches ProtScale.

What does a hydropathy value above 1.6 mean?

With a 19-residue window, a stretch averaging above 1.6 is hydrophobic enough to be a membrane-spanning helix. Kyte and Doolittle proposed the cutoff in their 1982 paper in the Journal of Molecular Biology, calibrated for their scale and a 19-residue window, so it does not apply to other scales or window sizes.

Why does the plot not start at residue 1?

Every point is the average of a full window centered on that residue. The first and last (window - 1) / 2 residues have no full window, so a 19-residue window starts at residue 10. This is how Kyte and Doolittle and ExPASy ProtScale plot hydropathy.

Is a hydropathy plot the same as a hydrophobicity plot?

A hydropathy plot is a hydrophobicity plot drawn with the Kyte-Doolittle hydropathy scale. The Hydrophobicity plot offers 24 other scales.

Is a positive value hydrophobic?

Yes. On the Kyte-Doolittle scale, positive values are hydrophobic and negative values are hydrophilic.

What is the difference between hydropathy and GRAVY?

Hydropathy is the Kyte-Doolittle value of a residue or the mean over a window, plotted along the chain. GRAVY is the same mean taken over the whole protein, so it gives one number per sequence and no positions.