- Add or delete the sections that you require.
Using statistical domain-domain interaction potentials to predict dimeric structures from sequences
Author(s): Sascha Willuweit, Patrick May
Affiliations: Institute of Legal Medicine, Deptartment for Forensic Genetics, Charite - Universitätsmedizin Berlin, Germany, Zuse Institute Berlin, Germany
Contact:email: email@example.com, firstname.lastname@example.org
Keywords: structural and computational genomics
Many biological processes such as signal transduction, transport, cellular motion and regulatory mechanisms are mediated by domain-domain interactions, thus the study of such interactions could gain our understanding of protein function. We have collected 90,166 distinct domain-domain interactions based on 36,837 Protein Databank (PDB) structures [1,2] together with their quaternary assembly given by the corresponding biounit files. Each domain was classified according to the SCOP  Version 1.69, allowing us to assign every domain-domain interaction to one of the 2,219 SCOP superfamily-superfamily clusters we identified. With this dataset we build several distinct residue-based statistical potentials [4,5] using the following variable interaction criterias: distance between heavy atoms of two residues, secondary structure and solvent exposure. These criteria combinations were evaluated in all common classes of interactions (heterodimer, homodimer, inter-chain and intra-chain) using a simple gapless threading algorithm, which should recognize the native interacting residues based on one of the potentials specified above. We found one statistical potential with appropriate parameter set to have the best results in all interaction classes: residues have to be exposed or intermediate solvent exposure with a distance less or equal to 4.0 Ängström and no restriction of secondary structure types. The simple gapless threading algorithm recognised the native interactions with 84.4% as the best, with 99.1% within the top five best and with 99.5% within the top ten best interactions. Additional experiments were performed with an extended version of our Branch-and-Bound-based protein threading core THESEUS [6,7], that was extended by additional constraints to take the contribution of the domain-domain interaction potential into account. THESEUS was able to predict 3D structures of several domain-domain interactions for selected examples.
Add your text here
Add your images or tables here
Add your text here
Add your text here
1. F. C. Bernstein, T. F. Koetzle, G. J. Williams, E. F. Meyer, M. D. Brice, J. R.
Rodgers, O. Kennard, T. Shimanouchi, and M. Tasumi. The Protein Data Bank: a
computer-based archival file for macromolecular structures. J Mol Biol, 112(3):535–
542, May 1977.
2. H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig, I. N. Shindyalov, and P. E. Bourne. The Protein Data Bank. Nucleic Acids Res, 28(1):235–242, Jan 2000.
3. A. G. Murzin, S. E. Brenner, T. Hubbard, and C. Chothia. SCOP: a structural classification of proteins database for the investigation of sequences and structures. J Mol Biol, 247(4):536–540, Apr 1995.
4. H. Lu and J. Skolnick. A distance-dependent atomic knowledge-based potential for improved protein structure selection. Proteins, 44(3):223–232, Aug 2001.
5. Hui Lu, Long Lu, and Jeffrey Skolnick. Development of unified statistical potentials describing protein-protein interactions. Biophys J, 84(3):1895–1901, Mar 2003.
6. P. May, T. Steinke, and M. Meyer. THESEUS: A parallel threading core. In Proceedings of the 12th International Conference on Intelligent Systems for Molecular Biology (ISMB) and 3rd European Conference on Computation Biology (ECCB), Glasgow, UK, 2004.
7. P. May and T. Steinke. THESEUS – protein structure prediction at ZIB. Technical report, Zuse Institute Berlin, 2006.