Hydrophobic and hydrophilic are useful starting words for describing how molecular groups behave around water. In a peptide, however, several kinds of groups are connected in one molecule. A sequence score compresses that chemistry and should be read with the scale and setting that produced it.
Water is part of the interaction
IUPAC describes hydrophobic interactions in terms of hydrocarbon-like groups associating in an aqueous setting and the accompanying water–water interactions. It discourages hydrophobic bond as a label. The phenomenon is not a new covalent connection between two side chains.IUPAC Gold Book — Hydrophobic interaction (opens in a new tab)
This matters when reading a structure illustration: two non-polar surfaces placed together are not chemically joined merely because the caption calls their contact hydrophobic. Nor should water be imagined as a passive background that has no role in the comparison.
Hydrophilic is also a tendency rather than a complete recipe. A group may interact favourably with water while the entire molecule has other features that affect its physical behaviour.
Ask what a numerical scale measures
Wimley and White derived an interfacial hydrophobicity scale using small model peptides partitioning into neutral phospholipid membrane interfaces. Its setting is part of its meaning. A membrane interface is not the same environment as bulk water or the interior of a hydrocarbon phase.Wimley and White — Interfacial hydrophobicity scale (opens in a new tab)
| Question | Why it matters |
|---|---|
| Which scale? | Different reference processes can produce different values |
| Which sign convention? | A larger number does not have one universal meaning |
| Which molecular groups? | Side-chain-only and whole-residue descriptions differ |
| Which environment? | An interface result is not automatically a solubility result |
A sensible report names the scale alongside the score. Copying a number without its definition makes apparently precise comparisons difficult to interpret and can reverse the intended ranking if the sign convention changes.
An average loses the sequence pattern
Consider two hypothetical six-residue strings built from three symbols P and three symbols N, where P means polar and N means non-polar for this illustration. PPPNNN and PNPNPN have identical counts. Any simple sum of the same six assigned values is therefore identical, but the order differs.
That arithmetic fact does not establish what either string folds into. It shows why an average alone cannot describe the spatial distribution of groups in a real conformation. Actual amino-acid identities and structure are needed before drawing a surface map.
This is also why a sequence can contain both polar and non-polar residues without necessarily presenting one continuous polar face and one continuous non-polar face. Composition and arrangement answer different questions.
Whole-peptide properties need whole-peptide evidence
Kuroda and colleagues compared equal-length homotetrapeptide systems in molecular simulations alongside experimental solubility observations. Their study is a concrete example of examining composition while holding length fixed. The simulations remain model results, not direct measurements of every possible sequence.Kuroda and colleagues — Multi-peptide simulations and solubility (opens in a new tab)
For a new peptide, a hydrophobicity score can motivate a prediction to test. It does not supply an experimentally established dissolved concentration, establish passage across a cell membrane or identify a biological effect.
- Name the scale and its environment.
- Retain ionisation and modifications where relevant.
- Distinguish an average score from a spatial pattern.
- Label measured and predicted properties separately.
Sources and further detail
- IUPAC Gold Book — Hydrophobic interaction (opens in a new tab)
The aqueous solvent is part of the phenomenon; hydrophobic bond is discouraged as a term.
- Wimley and White — Interfacial hydrophobicity scale (opens in a new tab)
Nature Structural Biology 3, 842–848 (1996), DOI 10.1038/nsb1096-842. Primary small-peptide partitioning study at neutral phospholipid interfaces; not a universal solubility scale.
- Kuroda and colleagues — Multi-peptide simulations and solubility (opens in a new tab)
Scientific Reports 6, 19479 (2016). Equal-length homotetrapeptide models allow composition effects to be distinguished from chain length; simulation and experimental comparisons are identified separately.
Sources checked 19 September 2026. Worked examples are illustrative unless a supplied report is explicitly identified. This article has not undergone independent scientific peer review.