Secondary structure
Assigns helix, strand and coil to every residue of a structure using DSSP, reported as contiguous elements like "helix, A12-A24". Use to describe a fold, compare two structures' architecture, or find which regions are ordered before trimming or modelling.
Secondary structure
What it does
Takes one structure and says which residues are in a helix, which are in a beta strand, and which are neither. The answer comes back as elements — "alpha helix, chain A, residues 12 to 24" — rather than one row per residue, because that is the form anyone actually reads.
The per-residue assignment is in the output files if you need it.
How the assignment is made
The engine is DSSP, and its definitions are the ones most papers mean when they say "helix" or "strand". It works from backbone hydrogen bonds: it calculates the electrostatic energy between each backbone N-H and every C=O, decides which pairs count as bonded, and then reads the pattern.
A helix is a run of residues each bonded to the one four along. A strand is a residue bonded to a partner in a different part of the chain. Everything else is turn, bend, or nothing.
Because it reads hydrogen bonds rather than shape, DSSP needs a complete backbone including carbonyl oxygens. A file with only alpha carbons has nothing for it to work with, and will come back with no assignments rather than a guess.
Reading the output
Eight states are reported, grouped three ways for the summary:
- helix — H (alpha), G (3-10), I (pi)
- sheet — E (strand), B (isolated bridge)
- coil — T (turn), S (bend), and unassigned
The summary gives the percentage of each. Above roughly 45% helix with little sheet is an all-alpha protein; above 30% sheet with little helix is all-beta; most things are mixed.
Each element also carries its chain and residue range, so you can quote it, trim to it, or compare it against another structure of the same protein.
Important limitations
It describes one file, not the molecule. A loop that is disordered in a crystal will be reported as coil, or missing entirely, whether or not it is structured in solution.
Short elements are not always meaningful. A two-residue "strand" is usually noise; the length is reported so you can judge.
It says nothing about function. Secondary structure is architecture. Two proteins with the same fold can do entirely different things.