Deciphering the structural code for proteins: Helical propensities in domain classes and statistical multiresidue information in α‐helices