A Study on the Covalent Geometry of Proteins and Its Impact on Distance Geometry

Abstract

The Distance Geometry Problem (DGP) involves determining the positions of a set of points in space based on the distances between them. Due to the importance of measuring spatial proximity’s between atoms in structural biology, Distance Geometry finds a natural application to protein structure determination. When an instance of the DGP is discretizable, the search tree can be explored using the (interval) branch-and-prune algorithm (iBP). A statistical study on a database of protein structures shows that slight variations are observed in the covalent geometry of the molecules, depending on the Ramachandran region of the protein residues. These variations are in agreement with observations made on high resolution X-ray crystallographic structures. These slight variations may cause atomic clashes when using iBP algorithms to calculate a folded protein structure. In this contribution, we compare two software implementations (MDjeep and ib-png) that attempt to address this issue.