Application of Mixture Design and Kohonen Neural Network for Determination of Macro- and Microelement in Mullet (Mugil cephalus) by MIP OES

• Autor(a) principal: Viana Pacheco, Clissiane Soares
• Data de Publicação: 2021
• Outros Autores: Costa, Floriatan Santos, Guedes, Wesley Nascimento [UNESP], Jesus, Marina Santos de, das Chagas, Thiago Pereira, Pinto dos Santos, Ana Maria, Lima, Daniel de Castro, Paranhos da Silva, Erik Galvao
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1007/s12161-021-01969-7; http://hdl.handle.net/11449/209115
• topic: Mugil cephalus; Mixture design; Kohonen self-organizing map; MIP OES

In this study, the concentrations of macro- and microelement in different tissues (liver, gill, and muscle) of fish collected (Mugil cephalus) in Pontal Bay (Ilheus, Bahia) were determined. A simplex-centroid mixture design was used to obtain the mixture applied to the decomposition of the samples. The elementary determination was made by microwave-induced plasma optical emission spectrometry (MIP OES), and the Kohonen self-organizing map (KSOM) was applied as an exploratory analysis tool. The accuracy of the procedure was assessed using certified reference material. Recoveries were obtained in the range of 99-103% and without a statistical difference (95% confidence level). Precision was estimated from the relative standard deviation, which ranged from 0.3 to 2.2%, and the limits of detection (mg kg(-1)) were 0.145 (Cu), 1.53 (Fe), 1.61 (K), 0.250 (Mg), 0.040 (Mn), 4.84 (P), and 2.12 (Zn). The KSOM provided efficient and precise separation of the studied tissues into three groups. The concentrations of Cu, Fe, and Zn were higher in the liver, while in the gills, they were the Mn.