Book of Abstracts: Albany 2003
June 17-21 2003
Pose Analysis of Structures of Alpha-Carbons in Proteins
We present a novel method to describe the relative pose (position and orientation) distribution of amino acid residue pairs within a protein that are proximal in space and distal in sequence. While the Ramachandran plot provides information of protein conformations using the φ and ψ angles between sequentially proximal residues, our method can offer six-dimensional relative pose information. Given a protein, we first affix a frame of reference to the alpha-carbon atom (Cα) of each amino acid in the structure. Then we record all possible positions and orientations between amino acids that are proximal in space and distal in sequence i.e., within certain spatial/sequential distance cutoffs. Hence, in essence we seek a six-dimensional Ramachandran-like plot for sequentially distant residue pairs. Since we deal with the data in a higher dimension than the two dimensional φ - ψ plane and handle a large amount of data taken from various proteins in the protein data bank (PDB), interpreting and visualizing the large amount of data effectively is a significant problem. Distribution data is visualized in the form of continuous distributions by using Gaussian distribution functions on SO(3) and R3. Hence, we discuss how the classical Gaussian functions can be generalized to capture both positional and orientational data.
The presented method was applied to 168 protein structures with the resolution of 2.0 Å or better, and R-factor less than 20% in the Protein Data Bank (PDB). Relative positional and orientational distribution plots were classified according to group types, i.e., hydrophobic, charged, polar, and characteristics for each group pair type (e.g., hydrophobic - hydrophobic) were discussed. Multiple clusters were found in many group types and sources of such clusters in distribution plots were also discussed. It was found that residues in secondary structures, i.e., helices or sheets, made significant contributions. For comparison, residue pairs that are both sequentially and spatially proximal were investigated. The mathematical techniques of pose analysis have been a useful tool to characterize the distribution of relative position and orientation of residue pairs.
Department of Mechanical Engineering