GRAVY

How to calculate a GRAVY score

Paste a protein sequence, either as plain one-letter amino acid text or in FASTA format, and run the calculator. ProteinIQ averages the Kyte-Doolittle hydropathy value across every residue and returns a GRAVY (Grand Average of Hydropathy) score rounded to three decimal places for each sequence. Use it as a compact description of overall protein hydrophobicity, not as a position-by-position membrane-topology prediction.

For example, this short sequence contains three strongly hydrophobic residues:

>AIV_reference
AIV

It returns a GRAVY score of 3.500, because (1.8 + 4.5 + 4.2) / 3 = 3.5.

The calculator also accepts several FASTA records in one submission. This input:

>AIV_reference
AIV
>charged_peptide
DEKR

returns these rows:

FASTA descriptions are kept separate from the sequence. The protein ID is the first space- or pipe-delimited token in a FASTA header. For example, >sp|P12345|NAME description is displayed as sp. Give records distinct identifiers when comparing several proteins.

For a biological example, the mature human insulin B and A chains can be compared separately:

>insulin-B human insulin B chain
FVNQHLCGSHLVEALYLVCGERGFFYTPKT
>insulin-A human insulin A chain
GIVEQCCTSICSLYQLENYCN

Input

The tool rejects a sequence containing a noncanonical residue such as B, J, O, U, X, or Z, or symbols such as alignment gaps, stop markers, punctuation, and position numbers. These characters are not silently removed to produce a score. Correct the input before rerunning; a submission containing an invalid or empty record returns an error rather than an incomplete comparison table.

Keep names and descriptions in FASTA headers. Plain text is interpreted as one sequence, so a short name containing only amino acid letters, such as GFP, is indistinguishable from a three-residue peptide. For a structure file, compare the observed-chain score with a full-length sequence score only after checking sequence coverage. An RCSB FASTA fetch uses the fetched sequence, which can differ from the residues resolved in the coordinates.

Settings

GRAVY has no user-configurable settings. It always uses the Kyte-Doolittle hydropathy scale and rounds the displayed score to three decimal places.

Results

The results table can be copied or downloaded for batch comparisons.

What does a GRAVY score mean?

GRAVY, the grand average of hydropathy (also written grand average of hydropathicity), is a whole-protein hydrophobicity index on the Kyte-Doolittle scale. It is calculated as:

Here, $N$ is the number of residues and $h_i$ is the Kyte-Doolittle hydropathy value assigned to residue $i$, using the published scale values and the averaging formula described by ExPASy ProtParam. Isoleucine (I, 4.5), valine (V, 4.2), leucine (L, 3.8), and phenylalanine (F, 2.8) raise the average. Charged residues such as arginine (R, -4.5), lysine (K, -3.9), aspartate (D, -3.5), and glutamate (E, -3.5) lower it.

A positive score means hydrophobic residues outweigh hydrophilic ones on this scale. A negative score means the reverse. The sign is more reliable than a universal cutoff: protein length, domain composition, oligomerization, disulfides, cofactors, and experimental conditions all affect real solubility and localization. A positive whole-protein average can support a membrane-association hypothesis, but it does not identify a transmembrane segment or establish that a protein is insoluble.

GRAVY also cannot distinguish where hydrophobic residues occur. Human glycophorin A crosses the red blood cell membrane with one helix, and its 19-residue hydropathy plot peaks at 2.67 over that helix, yet the full 150-residue sequence has a GRAVY of -0.080.

The polar extracellular and cytoplasmic parts outweigh the helix, so the local plot is the better test for membrane segments. For the same kind of whole-sequence mean on Eisenberg, Hopp-Woods, or another scale, the Hydrophobicity plot gives a peptide's mean when the window equals its length (odd lengths up to 21).

Which protein property tool should I use?

FAQ

Is GRAVY the same as a hydropathy plot?

No. GRAVY is one average over the complete sequence. A hydropathy plot keeps positional information by calculating local averages, which makes it more useful for inspecting candidate membrane-spanning segments.

Does a positive GRAVY score prove that a protein is membrane-bound?

No. It shows that the sequence is hydrophobic on average according to the Kyte-Doolittle scale. Confirming membrane association requires local sequence context and, where appropriate, structural or experimental evidence.

Why was my protein sequence rejected?

Only the 20 standard amino acid letters have supported values in this calculator. Check for ambiguity codes, modified residues, gaps, stop markers, position numbers, or nonprotein text. Every FASTA header also needs sequence data on the next line. Resolve ambiguous residues from the source sequence rather than choosing a replacement arbitrarily.

Can I calculate GRAVY for multiple FASTA sequences?

Yes. Include each sequence as a separate FASTA record. The result table contains one row per validated sequence, with its identifier, length, and GRAVY score.

Is GRAVY a hydrophobicity index?

Yes. It is the mean Kyte-Doolittle hydropathy index of all residues, so it describes the whole sequence on that one scale. Other scales give different numbers for the same protein, so compare GRAVY values only with other GRAVY values.