Desenvolvimento e validação de novos métodos de distribuição da população inicial em algoritmos genéticos para o problema de docking proteína-ligante

• Ano de defesa: 2008
• Autor(a) principal: Gonçalves, Reinaldo Bellini
• Orientador(a): Barbosa, Helio José Corrêa
• Banca de defesa: Borges, Carlos Cristiano H. , Caffarena, Ernesto Raúl
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Laboratório Nacional de Computação Científica
• Programa de Pós-Graduação: Programa de Pós-Graduação em Modelagem Computacional
• Departamento: Serviço de Análise e Apoio a Formação de Recursos Humanos
• País: BR
• Palavras-chave em Português: Algoritmos genéticos; Biologia molecular - Processamento eletrônico de dados; Inteligência artificial - Aplicações biológicas; Matemática aplicada; Desenho racional de fármacos
• Palavras-chave em Inglês: Genetic algorithms; Artificial inteligence
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::PROBABILIDADE::PROCESSOS ESTOCASTICOS ESPECIAIS
• Link de acesso: https://tede.lncc.br/handle/tede/84

The methods of protein-ligand docking are computational methods usedto predict the mode of binding of molecules into drug candidates for its receptor. The docking allows tests of hundreds of compounds in ashort space of time, assisting in the discovery of new drug candidates. The great complexity that involves the binding of protein-ligand complex, makes the problem of docking computationally difficult to be solved. In this work, we used the Genetic Algorithms which is a technique of optimization based on the theory of biological evolution of Darwin. The proposed algorithm was implemented and tested initially by Camila S. de Magalhães in her doctoral thesis, with the Group of Molecular Modeling of Biological Systems at LNCC, with a range of 5 ligands of HIV-1 protease. It was built a new set used for test with 49 structures with several physico-chemical properties, distributed in 22 different families of protein, allowing for a broader test of the algorithm It was conducted a detailed study of the dependence of the genetic algorithm in relation to the distribution of its initial population and it was also investigated ways more efficient and robust to generate the same. Among these, the proposal to distribute the initial population based on the coordinates of individuals of lower energy in the population (proposal 5), it is very promising. This distribution has allowed the algorithm to obtain good results, finding solutions of lower energy in the population very close to experimental structure optimized, without having specific information about the experimental structure. This fact is very important, because the algorithm makes it more realistic in view that in the rational design of drugs, it has not the trial structure.