Molecular networks by invariant shape coordinates and deformation indexes

Abstract

The classification of large molecules according to structural similarities is a relevant issue in biochemistry [1], even more in the upcoming big data era. The proper choice of parameters, containing invariant structural information, is a convenient way to induce mapping and grouping of structures, depending on predominant structural motifs, individual amino acid geometry or connectivity properties. Few-body hyperspherical coordinates provide the appropriate framework in which shape parameters and deformation indices [2-4] may be defined and applied to characterize complex structures.