Estudo teórico (modelagem molecular e QSAR) de compostos quinolínicos com atividade herbicida

• Ano de defesa: 2017
• Autor(a) principal: Ribeiro, Taisa Pereira Piacentini
• Orientador(a): Melo, Eduardo Borges de
• Banca de defesa: Melo, Eduardo Borges de , Faria , Terezinha de Jesus , Rosa, Mauricio Ferreira da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual do Oeste do Paraná; Cascavel
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciências Farmacêuticas
• Departamento: Centro de Ciências Médicas e Farmacêuticas
• País: Brasil
• Palavras-chave em Português: Inibição do PET; Estudos de relação estrutura-atividade quantitativa; Regra de três de Briggs; Triagem virtual
• Palavras-chave em Inglês: PET-inhibiting; Studies of quantitative structure-activity relationship; Briggs rule of three; Virtual screening
• Área do conhecimento CNPq: CIENCIAS DA SAUDE::FARMACIA
• Link de acesso: http://tede.unioeste.br/handle/tede/2964

The search for new herbicides to control herbicides-resistant weeds is necessary to attend the rising demand of food from the world’s population. This work was divided into two parts. The first aimed to obtain a model of QSAR-2D, 3D and hybrid to predict compounds with activity to the inhibition of photosynthesis. For this, was used a data set of 44 quinoline analogues described in the literature as PET inhibitors, and all tested in the same bioassay method. For construction of models were used the programs QSAR Modeling and Pentacle. The obtained models A, C and D, were approved in the validation tests (internal and external), they are robust and with good predictive capacity. The second part of studie aimed to identify a pharmacophore model, for select compounds from the data set of first part, aiming to use as a tool for virtual screening. The research resulted in 86,560 compounds, and thus several screening filters were applied according to Briggs rule of three, in silico toxicity analyzes, unsupervised pattern recognition (PCA), and docking studies. As a result, 28 compounds remained, all of which showed potential to be herbicides, through the prediction using the obtained QSAR models, however, only the model D proved to be reliable for prediction the virtual screening. Finally, we selected the ten compounds that presented the highest predictive value of PET inhibition activity, using the model D.