 |
    |
Computer-Aided Peptide Evolution for Virtual Drug Design
Ignasi Belda1, Xavier Llorà2, Marc Martinell1, Teresa Tarragó1, and Ernest Giralt1,3
1Institut de Recerca Biomèdica de Barcelona, Parc Científic de Barcelona, Universitat de Barcelona, E 08028 Barcelona, Spain.
ibelda@pcb.ub.es
mmartinell@pcb.ub.es
ttarrago@pcb.ub.es
egiralt@pcb.ub.es
2Illinois Genetic Algorithms Laboratory, National Center for Supercomputing Application, University of Illinois at Urbana-Champaign, Urbana, IL 61801
xllora@illigal.ge.uiuc.edu
3Departament de Química Orgànica, Universitat de Barcelona., E 08028 Barcelona, Spain.
Abstract. One of the goals of computational chemistry is the automated de novo design of bioactive molecules. Despite significant progress in computational approaches to ligand design and efficient evaluation of binding energy, novel procedures for ligand design are required. Evolutionary computation provides a new approach to this design issue. A reliable framework for obtaining ligands via evolutionary algorithms has been implemented. It provides an automatic tool for peptide de novo design, based on protein surface patches defined by user. A special emphasis has been given to the evaluation of the proposed peptides. Hence, we have devised two different evaluation heuristics to carry out this task. Then, we have tested the proposed framework in the design of ligands for the protein Prolyl oligopetidase, p53, and DNA Gyrase.
LNCS 3102, p. 321 ff.
lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2004