Identification of resistant sugarcane genotypes to ratoon stunt disease by Fourier-transform infrared analysis

• Ano de defesa: 2019
• Autor(a) principal: Laborde, Marie Caroline Ferreira
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Leifsonia xyli subsp. xyli; Saccharum officinarum; Infrared spectroscopy; Melhoramento da cana-de-açúcar; Sugarcane breeding
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/11/11135/tde-14082019-104403/

Ratoon stunting disease (RSD), caused by Leifsonia xyli subsp. xyli (Lxx) can cause substantial impact on biomass production in all sugarcane producing countries. Few management options exist for controlling the disease, and they are all centered on measures of hygiene and sanitation. These measures are not efficient in eliminating the bacteria and can damage bud germination. Phenotypic selection would be the best management practice. However, selection based on Lxx-restrictive multiplication is costly, time-consuming, and destructive. One potentially useful approach for identifying resistant genotypes is Fourier-transform infrared (FT-IR) spectroscopy. In this study, we demonstrate that by using qPCR to quantify bacterial titers and to determine the resistance levels of nineteen commercial varieties inoculated with Lxx and FT-IR spectroscopy we could identify resistant and susceptible genotypes based on specific functional groups. Using a non-linear machine learning techniques (SVM) and a linear partial least square regression (PLSR) we distinguished between resistant and susceptible sugarcane genotypes and predict bacteria concentration using spectral data and Lxx titer quantification. Our Results suggest that resistance may be associated with variation in chemistry. The resistant mechanisms involved against RSD are probably associated to polyamines regulation and aromatic rings compounds accumulation. The results demonstrated that both approaches, SVM and PLSR, are capable of providing accurate estimates of sugarcane resistance to Lxx and can be applied to screen resistant sugarcane in breeding programs.