Protein structures are classically described in terms of secondary structures, i.e., two regular states, the alpha-helices and the beta-strands and one default state, the coil. Even if the regular secondary structures have relevant physical meaning, the definition of secondary structures has some important (and often forgotten) limitations: the rules for secondary structure assignments are (i) not simple, (ii) not unique and (iii) 50% of all residues, which occur in the coil, are not described. Hence, different research groups have described local protein structures with the aim of analyzing them and to approximate every part of the protein backbone. These libraries of local structures consist of sets of small prototypes named "structural alphabets". They have also been used to predict the protein backbone conformation. In this chapter, we first present the secondary structures, i.e., the most classical approach to describe protein structures, followed by the different structural alphabets designed till date. We focus on the different prediction schemes developed with these structural alphabets.