Análise exploratória e preditiva de linhagens de caupí baseada nas respostas oxidativas à presença de alumínio

• Autor(a) principal: Oliveira, Uasley Caldas de
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UEFS
• Texto Completo: http://tede2.uefs.br:8080/handle/tede/1651
• topic: Feijão-caupí; Vigna unguiculata (L.) Walp.; Alumínio; Concentração; Modelagem preditiva; Estresse abiótico; Catalase; Espécies reativas de oxigênio; Aprendizado de máquina; Vigna unguiculata (L.) Walp.; Abiotic stress; Machine learning; Reactive oxygen species; Random Forest; Aluminum; Concentration; CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA; GENETICA::GENETICA VEGETAL; FITOTECNIA::PRODUCAO E BENEFICIAMENTO DE SEMENTES; CIENCIAS BIOLOGICAS::GENETICA

Aluminum toxicity is a significant limiting factor in cowpea (Vigna unguiculata L.) cultivation, a legume grown in tropical and subtropical regions, highly valued as a food source, particularly for its nutritional and socio-economic importance to small-scale farmers in the North and Northeast regions of Brazil. The objective of this study was to assess cowpea lineages for aluminum ion tolerance based on the specific activity of antioxidant enzymes using predictive modeling. The experiment was conducted at the State University of Feira de Santana (UEFS), in the Seed Germination Laboratory (LAGER), and in a greenhouse. Evaluation of protein content and the specific activity of the enzymes ascorbate peroxidase (APx), catalase (CAT), peroxidase (POD), and superoxide dismutase (SOD) was performed on plants exposed to different aluminum concentrations. Various forms of aluminum tolerance classification were used to analyze the results, including mean tests using Scott-Knott with a probability level of p<0.05. Additionally, predictive modeling was employed, incorporating data trees, where predictive models like Random Forest, Tree, Neural Network, and kNN were tested. Evans blue dye was utilized as a visual indicator of aluminum toxicity, and its quantification was done through spectrophotometry. Significant genetic variability was observed among cowpea lineages concerning aluminum tolerance. The enzyme activity data from plants exposed to aluminum ions enabled the determination that the Random Forest and Neural Network models for images with Evans Blue dye displayed the best predictive capability for both data sets. Using the visual method with Evans blue dye, lineage O demonstrated the highest tolerance, while lineages I, J, F, G, and M were the most sensitive to aluminum, as determined by root apex coloration. Regarding Evans blue quantification, lineages D and C were the most tolerant, and the most sensitive lineages were G and A.