Abstract
BackgroundAligning similar molecular structures is an important step in the process of bio-molecular structure and function analysis. Molecular surfaces are simple representations of molecular structure that are easily constructed from various forms of molecular data such as 3D atomic coordinates (PDB) and Electron Microscopy (EM) data.MethodsWe present a Multi-Scale Morse-Smale Molecular-Surface Alignment tool, MS3ALIGN, which aligns molecular surfaces based on significant protrusions on the molecular surface. The input is a pair of molecular surfaces represented as triangle meshes. A key advantage of MS3ALIGN is computational efficiency that is achieved because it processes only a few carefully chosen protrusions on the molecular surface. Furthermore, the alignments are partial in nature and therefore allows for inexact surfaces to be aligned.ResultsThe method is evaluated in four settings. First, we establish performance using known alignments with varying overlap and noise values. Second, we compare the method with SurfComp, an existing surface alignment method. We show that we are able to determine alignments reported by SurfComp, as well as report relevant alignments not found by SurfComp. Third, we validate the ability of MS3ALIGN to determine alignments in the case of structurally dissimilar binding sites. Fourth, we demonstrate the ability of MS3ALIGN to align iso-surfaces derived from cryo-electron microscopy scans.ConclusionsWe have presented an algorithm that aligns Molecular Surfaces based on the topology of surface curvature. A webserver and standalone software implementation of the algorithm available at http://vgl.serc.iisc.ernet.in/ms3align.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0874-8) contains supplementary material, which is available to authorized users.
Highlights
Three dimensional solved crystal structures of proteins provide valuable insights regarding the function of the protein as the precise position of all functionally and structurally important residues is known
In the first experiment (“Performance analysis” section), we evaluate its performance under conditions of noise, partial overlap, and running times using a random set of 20 proteins from the Protein data bank (PDB) [31]
We present MS3ALIGN, a tool that implements this method
Summary
Three dimensional solved crystal structures of proteins provide valuable insights regarding the function of the protein as the precise position of all functionally and structurally important residues is known. The function of an unknown protein may be determined by comparing its structure to structures of proteins whose functions are already known Tools such as MUSTANG and DALI [1, 2], which are widely used to compare protein structures, use three-dimensional co-ordinates of atoms in the protein structures as inputs and report structural dissimilarities in terms of an RMS distance between their aligned coordinates. Proteins that function as enzymes and transporters contain a pocket or the binding site in the structure that accommodates the substrate and cargo small molecules respectively. Pocket and ligand alignments could potentially provide insights into protein function. The curvature at a point t is defined as the magnitude of the second differential at t, κ (t) d2l dt
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call