Efficient optimization of electrostatic interactions between biomolecules.
We present a PDE-constrained approach to optimizing the electrostatic interactions between two biomolecules. These interactions play important roles in the determination of binding affinity and specificity, and are therefore of significant interest when designing a ligand molecule to bind tightly to a receptor. Using a popular continuum model and physically reasonable assumptions, the electrostatic component of the binding free energy is a convex, quadratic function of the ligand charge distribution. Traditional optimization methods require exhaustive pre-computation, and the expense has precluded a full exploration of the promise of electrostatic optimization in biomolecule analysis and design. In this paper we describe an approach in which the electrostatic simulations and optimization problem are solved simultaneously; unlike many PDE- constrained optimization frameworks, the proposed method does not incorporate the PDE as a set of equality constraints. This co-optimization approach can be used by itself to solve unconstrained problems or those with linear equality constraints, or in conjunction with primal-dual interior point methods to solve problems with inequality constraints. Model problems demonstrate that the co-optimization method is computationally efficient and can be used to solve realistic problems.
- Research Organization:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC); National Institutes of Health (NIH); National Science Foundation (NSF); Singapore-MIT Alliance
- DOE Contract Number:
- DE-AC02-06CH11357
- OSTI ID:
- 971155
- Report Number(s):
- ANL/MCS/CP-59820; TRN: US201003%%602
- Resource Relation:
- Conference: 46th IEEE Conference on Decision and Control; Dec. 12, 2007 - Dec. 14, 2007; New Orleans, LA
- Country of Publication:
- United States
- Language:
- ENGLISH
Similar Records
Quantum Annealing with Inequality Constraints: The Set Cover Problem
An infeasible-start framework for convex quadratic optimization, with application to constraint-reduced interior-point and other methods