Aplicação de espectrometria de massas com ionização por eletrospray e métodos quimiométricos para classificar espécies de Miconia spp.

• Ano de defesa: 2023
• Autor(a) principal: Achuate, Assumane Joaquim
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Química
• Palavras-chave em Português: Gênero Miconia; Genus Miconia; ESI-MS; Quimiometria; Chemometrics; PLS-DA; DD-SIMCA; CNPQ::CIENCIAS EXATAS E DA TERRA; Química; ODS::ODS 9. Indústria, Inovação e infraestrutura - Construir infraestrutura resiliente, promover a industrialização inclusiva e sustentável, e fomentar a inovação.
• Link de acesso: https://repositorio.ufu.br/handle/123456789/39960; http://doi.org/10.14393/ufu.di.2023.640

Because they have therapeutic properties, medicinal plants are used to treat, cure and prevent diseases, and this clinical practice is considered one of the oldest in humanity. Amongst the plants with medicinal properties, the genus Miconia is notable. This genus stands out for being one of the largest groups in the Melastomataceae family and for being distributed throughout the American continent. In Brazil, the genus Miconia is found in almost all plant formations and is represented by 26 species in Minas Gerais, where four of these species, namely M. albicans, M. chamissois, M. cuspidata and M. rubiginosa were studied in this work. This study aims to develop analytical methods capable of classifying the 4 species of Miconia using electrospray ionization mass spectrometry (ESI-MS) and chemometric methods. The use of Partial Least Squares Discriminant Analysis (PLS-DA) and Data Driven Smooth Independent Modelling of Class Analogy (DD-SIMCA) chemometric methods to extract the information in the mass spectra enabled the development of stable, robust and easy-to-interpret models. The use of the mass spectrometry technique in combination with chemometric methods to classify the samples of the 4 species proved to be efficient, since the PLS-DA and DD-SIMCA models built were able to correctly discriminate all the samples of the Miconia species in their respective classifications, obtaining an efficiency of 100% and sensitivity and specificity values equal to 1. The PLS-DA models developed have Root Mean Square Errors of Calibration (RMSEC), Cross Validation (RMSECV) and Prediction (RMSEP) below 1, demonstrating good accuracy. In this way, the analytical methods developed could become a viable alternative for quality control of natural products, and could be used for drug development and treatment of diseases.