A comparative molecular field analysis study on several bioactive peptides using the alignment rules derived from identification of commonly exposed groups

Abstract

A 3D convex hull computation algorithm designed by us previously is used as a tool to align a series of structures randomly generated for eleven bioactive tachykinin peptides. There are 10 random structures generated for each peptide. A random structure is selected for each peptide to form a structural set of eleven structures. The total number of structural sets generated is 100. The convex hull computation algorithm is applied to each peptide structure generated. We count the frequency of atoms lying on the vertices of each hull computed. Vertices of the same atom type are gathered together as a set of commonly exposed atoms for a structural set generated. Structures are then aligned by treating the set of commonly exposed atoms as a set of correspondences using the FIT option of the SYBYL 6.4 program. All the structure sets are also aligned by using the coordinates of the backbone C α atoms as a set of correspondences through the same program. It is found that while a smaller degree of structural similarity for structures aligned by the convex hull alignment rule is detected, the overall SYBYL comparative molecular field analysis (CoMFA) statistics computed for the aligned structures using the alignment rule is better than that computed for the aligned structures using the C α atoms alignment rule. A similar conclusion is drawn for a subset of structures selected and probed by a different type of atoms using the SYBYL CoMFA program. These results indicate that computation of 3D convex hulls is a feasible way that one can use to align structures generated for highly flexible molecules of this kind.