Abstract

GARDUNO JUAREZ, Ramón  and  MORALES, Luis B.. Memorias conformacionales en la predicción de estructura terciaria de polipéptidos. Rev. Soc. Quím. Méx [online]. 2003, vol.47, n.1, pp.6-21. ISSN 0583-7693.

To reduce the conformational space existing in the process of searching for the tertiary structure of polypeptides, which grows exponentially with the number of residues, we have designed a heuristic search algorithm that is robust and that can provide with structures near the native with ease. Specifically, we have developed a hybrid genetic algorithm (HGA) that uses real numbers instead of bits in order to describe the genes of any given chromosome, and that includes an improvement operator which function is to reorient the search towards the best fitted individuals. At the end of each HGA cycle, an analysis of the resulting population shows the presence of well defined dihedral angle clusters for each chromosome. These segments correspond to different low energy conformations, and are relatively constant each time a new HGA experiment is performed on the same molecule. We have called "conformational memories" to these segments, which are used as a reduced conformational space in which the new round of the HGA is performed. Use of conformational memories speeds up and refines the localization of the structure at the Global Energy Minimum (GEM). This algorithm has been used to predict successfully the GEM for Met- and Leu-enkephalin.

Keywords : Protein structure prediction; heuristic algorithms; genetic algorithms; protein folding; conformational space search; conformational memories.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )