Treating the full protein structure is often neither computationally nor physically possible. Instead one is forced to consider various reduced models capturing the properties of interest. Previous work have used tubular neighborhoods of the C-alpha backbone. However, assigning a unique radius might not correctly capture volume exclusion - of crucial importance when trying to understand a protein's 3d-structure. We propose a new reduced model treating the protein as a non-uniform tube with a radius reflecting the positions of atoms. The tube representation is well suited considering X-ray crystallographic resolution ~ 3Å while a varying radius accounts for the different sizes of side chains. Such a non-uniform tube better captures the protein geometry and has numerous applications in structural/computational biology from the classification of protein structures to sequence-structure prediction.
|Published - 2005
|1st Annual Symposium on Computational and Systems Biology - Institute of Bioinformatics, University of Georgia, Athens, United States
Duration: 1 Jan 2005 → …
Conference number: 1
|1st Annual Symposium on Computational and Systems Biology
|Institute of Bioinformatics, University of Georgia
|01/01/2005 → …