‘Evolutionary algorithms for de novo peptide design’

Ernet Giralt Lledó, Francecs Xavier Llorà i Fàbrega, Ignacio Belda i Reig, Marcelo Javier Kogan Bocian, Sergio Madurga Díez.

Evolutionary computation provides a new approach to this design issue. The aim of the project is to develop a reliable framework for obtaining ligands via evolutionary algorithms using very recent techniques from artificial intelligence.This framework will focus on providing an automatic tool for de novo peptide design, based on designs of protein peptide ligands. Accurate and fast methods are required to evaluate the quality of the peptides proposed by the evolutionary algorithms, and we have accordingly devised an accurate and efficient evaluation heuristic to carry out this task.The research team has already developed a preliminary approach to the proposed framework and has tested it in the design of ligands for the protein POP, p53 and DNA Gyrase. The preliminary results obtained in this real world domain are encouraging, suggesting great potential for the proposed framework.