Identification of convexity as a common structure feature for structures generated for two short peptides

Abstract

This paper presents a simple 3D convex hull computational algorithm for detecting some common structural features in structures generated by a molecular dynamics simulation for two short flexible tachykinin peptides. In this algorithm, each atom in each generated structure is treated as a point in space. First, a triangulating procedure is used to triangulate a set of points (a structure). Then, the volume of a series of corresponding parallelepipeds formed for each of a triangular facet generated by the arbitrary selection of a fourth point from among the other points of the set is sorted. The frequency of atoms lying on the vertices of a series of hulls that have been computed for a group of structures generated with the same structural topology are analyzed by both a cluster and a principal components analytical method. We find that about 15% of the total atoms commonly lie on the vertices of the two series of hulls computed for structures generated with two different starting structures but which have the same structural topology. However, for structures generated with two different structural topologies, the possibility of the presence of these common vertices is only about 5%. When detectable, these common structural features may be treated as a set of correspondences in a program such as CoMFA for aligning structures generated for the two highly flexible peptides.