Continuous Optimization Method for the Sequence Design of Protein Models Consisting of Multiple Monomer Types

, , ,

Keywords: , ,

In our earlier work, we had introduced continuous modeling of two discrete states (H: Hydrophobic; P: Polar) of the amino acid monomers and presented a deterministic continuous optimization method to design the sequence that minimizes the energy for a desired conformation of proteins. In this paper, we extend our previous method to handle multiple monomer types so that real proteins with more amino acid monomers than just HP could be designed with significant computational efficiency over the existing non-deterministic methods. By using two variables for each residue site (thus, 2N variables for an N-residue protein), continuous states of m different monomer types and the corresponding inter-residue interaction energy are defined. The energy is then minimized with respect to the 2N variables using efficient gradient-based optimization methods. The results with small-sized lattice models are used to validate the method. The initial guess for the optimization problem is generated in a preceding stage either by our previously reported quadratic programming formulation or by a statistical theory based method. The example of a real protein (PDB code: 1SRL) took less than 5 minutes on a desktop computer. The obtained minimum energy given for this case is lower than that of competing methods.

PDF of paper:

Journal: TechConnect Briefs
Volume: 1, Technical Proceedings of the 2005 NSTI Nanotechnology Conference and Trade Show, Volume 1
Published: May 8, 2005
Pages: 520 - 523
Industry sectors: Advanced Materials & Manufacturing | Medical & Biotech
Topics: Biomaterials, Informatics, Modeling & Simulation
ISBN: 0-9767985-0-6