Novel formulations for the sequence selection problem in de novo protein design with flexible templates

  • Authors:

  • H. K. Fung;M. S. Taylor;C. A. Floudas

  • Affiliations:

  • Department of Chemical Engineering, Princeton University, Princeton, NJ, USA;Department of Chemical Engineering, Princeton University, Princeton, NJ, USA;Department of Chemical Engineering, Princeton University, Princeton, NJ, USA

  • Venue:
  • Optimization Methods & Software - Systems Analysis, Optimization and Data Mining in Biomedicine
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents two novel formulations for solving the sequence selection problem in de novo protein design with highly flexible templates, each of which exhibits a crystal or NMR, structure. The first formulation applies weighted average energy parameters to incorporate information about every structure, with the weights, which are parameters dependent on a pair of Cα positions and a particular distance bin, given by the probability that the distance between the two positions is found to belong to that distance bin in any of the structures. The second formulation allows the distance between the two positions considered to fall into any distance bin that all the structures span over, but imposes novel linear constraints to ensure a physically consistent structure. Both formulations were tested on redesigning Compstatin, the template of which has 21 NMR structures from the protein data bank.