Edible seeds clustering based on phenolics and antioxidant activity using multivariate analysis

• Autor(a) principal: Barros, Hanna Elisia Araújo de
• Data de Publicação: 2021
• Outros Autores: Alexandre, Ana Cláudia Silveira, Campolina, Gabriela Aguiar, Alvarenga, Gabriela Fontes, Silva, Lara Maria dos Santos Ferraz e, Natarelli, Caio Vinicius Lima, Carvalho, Elisângela Elena Nunes, Vilas Boas, Eduardo Valério de Barros
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/49954
• topic: Bioactive compounds; Nuts; Principal component analysis; Hierarchical clusters analysis; Artificial neural network; Compostos bioativos; Nozes; Análise de Componentes Principais; Métodos hierárquicos da análise de cluster; Rede neural artificial

Edible seeds, especially those known by the population as nuts, have their consumption associated with functional appeal. The present study aimed to compare and group nine different seeds, traditional and regional, according to their similarities, in terms of moisture, total phenolic compounds (TPC) and antioxidant activity, through multivariate analyses. All results were submitted to Principal Component Analysis (PCA), Hierarchical Clusters (HCA) and Kohonen's self-organizing maps (ANN/KSOM). The seeds differed in terms of moisture content, TPC and antioxidant activity. The walnut butterfly stood out with the highest levels of TPC and antioxidant activity. In the multivariate analyses application, three groups were formed: i) hazel, baru, Brazil, macadamia, almond and cashew; ii) pequi and marolo; iii) walnut butterfly. It is concluded that the seeds can be separated into three groups, with ANN/KSOMs being the most self-explanatory analysis and that regional seeds are nutritionally similar to those traditionally consumed.