Figure 2. Schematic representation of the construction of family specific PEN at an energy cutoff ‘e’ and commonality coefficient ‘cc’ (f–PEN_e(cc)).

The steps are indicated with a simple example of two β–loop–β structural motifs, one structure with a short loop (structure 1) and another with a long loop (structure 2). (A) The PEN_es of the structures (1 and 2) are generated by connecting the residues based on their Cα–Cα distances (however the cutoff energy values (e) are chosen in the real cases to draw edges). (B) The structures are superimposed using MUSTANG [51]. (C) The structure based sequence alignment (MSSA) is obtained where the strands are aligned forming a set of equivalent nodes (VLKY and LCIKV) and the non–aligned loops are compensated using gaps in the MSSA. (D) Remapping of PEN_es of structures 1 and 2 on matrices of the same size (27×27) in which the gaps are represented as virtual nodes (VN, highlighted using self-edges). The arrays of nodes in both the structure networks (red and blue) are equivalent (i.e. Y31 (position 1st row and 2nd column) of structure 1 is structurally equivalent to Y12 of structure 2). (E) The f–PEN_e is obtained by aligning both the remapped PEN_e and edges are introduced in the network if they are present in any of the remapped PENs. (F) In this specific case the cc = 1.0 (i e. X = 2), and the family specific network represents only edges that are common to both the structures in the MSSA. The residues involved in the interactions in f–PEN_e(1.0) are highlighted as green spheres and the matrix of size 10×10 below the cartoon represents the interactions (X = 2) among the highlighted residues in both the structures.