What the complexity score measures

Each protein chain gets a composite score: a weighted blend of the aspects below, multiplied by a small-fold size gate so a tiny tidy domain can't be pushed high by intensive per-residue terms. The scale is open-ended — ≈0 for an unstructured peptide, ≈50 for a typical single domain, 90–100 for rare intricate folds, and above 100 for exceptionally complex multi-domain architectures. Peptides sit lowest (cyclic ones a little above linear), then small domains, then large multi-domain folds. The weights are re-calibrated from the votes and ratings on the other tabs.

Secondary structure is assigned with a built-in Kabsch–Sander (DSSP) hydrogen-bond algorithm; disorder and linkers are read out separately (ordered loops & structured inter-domain linkers add complexity, floppy unstructured tails do not).

Suggest an aspect we're missing

Is there a structural feature you think should count toward complexity — or one that's over- or under-weighted? Tell us and it goes straight to the team.