Eigenmode Analysis of Enzyme Carbon Backbones
I tried to assess the common properties of protein active sites with the eigenmode analysis of carbon backbones under Dr. Seonggu Ro at CrystalGenomics, Inc. from July to August 2002 (used languages: C++, Python). Given the three-dimensional information of a peptide, I simplified the peptide by ignoring the structures of side chains and transferring the mass of each side chain to the corresponding alpha carbon atom, leaving an (bead-like) alpha carbon backbone only. I assumed those atoms with different masses are connected by same springs (i.e., connected harmonic oscillators) and do not exhibit a conformational change (i.e., each carbon atom vibrates with respect to its fixed vibrational center). I inspected eigenmodes of the carbons in hopes of finding some common features of the carbon atoms of active sites, but I couldn’t find any significant results. I concluded that eigenmode information is not a good descriptor for active sites (or at least for the active sites of those simplified carbon backbones).
