Modeling the incidence of citrus canker in leaves of the sweet orange variety ‘Pera’

• Autor(a) principal: Pompeu, Danielle da Silva
• Data de Publicação: 2017
• Outros Autores: Guedes, Terezinha Aparecida, Janeiro, Vanderly, Gonçalves-Zuliane, Aline Maria Orbolato, Nunes, William Mario de Carvalho, Carneiro, José Walter Pedroza
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Acta Scientiarum. Agronomy (Online)
• Texto Completo: http://www.periodicos.uem.br/ojs/index.php/ActaSciAgron/article/view/28535
• topic: zero-inflated Beta distribution; mixture models; inflated Beta regression model; modeling proportions; Fitossanidade; Probabilidade e Estatística Aplicadas

Citrus canker, caused by the bacterium Xanthomonas citri subsp. citri, is one of the most important diseases of citrus. The use of resistant genotypes plays an important role in the management and control of the disease and is the most environmentally sustainable approach to disease control. Citrus canker incidence was recorded in an experiment on nine genotypes of the sweet orange variety ‘Pera’ grafted on four rootstocks. The experiment was started in 2010 and the incidence of citrus canker on the leaves was recorded on a quarterly basis. The incidence data from the experiment were analyzed using a zero-inflated Beta regression model (RBIZ), which is the appropriate method to describe data with large numbers of zeros. Based on the residual analysis, the data fit the model well. The discrete component of the explanatory variable, rootstock, was not significant as a factor affecting the onset of disease, in contrast with the continuous component, genotype, which was significant in explaining the incidence of citrus canker.