Secondary structure is assigned to every accepted conformation in order to maintain a structural propensity list, which is needed to determine each residue's conformational bias. Secondary structure assignments are based solely on backbone torsion angles. For this purpose, φ, ψ space is partitioned into 36 bins, each represented by a one-letter code as show here.

We refer to these bins as mesostates and to the one-letter codes as mesostate codes. To assign secondary structure, φ, ψ-values for each residue are mapped onto the corresponding mesostate code. In turn, the mesostate codes are mapped into a secondary structure class. The default value for each residue is initialized to coil, with mesostate code-->secondary structure mappings in precedence order as follows: PII is assigned to any residue in M or R; a turn is assigned to any consecutive pair from the set {OO, OP, OJ, PO, PP, PJ, JO, JP, JJ, Mo, Mp, Mj, Ro, Rp, Rj, oo, op, oj, po, pp, pj, jo, jp, jj, mO, mP, mJ, rO, rP, rJ}; helix is assigned to 5 or more repetitions of O or P; strand is assigned to 3 or more repetitions of L or G.

Note on ψ in the LINUS code:

The Z-matrix description of the protein makes it more efficient to describe ψ as the angle

N(i) - CA(i) - C(i) - O(i)

instead of

N(i) - CA(i) - C(i) - N(i+1)

Consequently the internal "psi" in LINUS working code is off by 180 deg. to the standard definition of ψ. However, all ψ values reported are corrected.