ABSTRACT Recently, the Capsicum chinense Jacquin pepper has been gaining prominence in the production of sauces, pastes and other food products. Due to the high demand for these peppers, it becomes even more necessary to obtain quality seeds to obtain greater productivity. Several factors can influence seed quality, so using logistic regression models to predict the percentage of seed germination can bring important results about the best production conditions. The present work aimed to compare logistic regression models using the ROC curve methodology to evaluate the germination capacity of Capsicum chinense Jacquin seeds. The factors evaluated were: variety (orange and red), fruit storage period (1, 7, and 14 days), extraction method (manual and mechanical), and seed storage period (3, 6, 9, and 12 months). By Akaike's information criterion (AIC) and the area under the ROC curve, the best model was the one with interaction between “cultivate”, “seed storage”, and “fruit storage”, without the “extraction method” factor. The model allowed to determine the chance of germination from the combination of these factors.

RESUMO Recentemente, a pimenta Capsicum chinense Jacquin tem adquirido destaque na produção de molhos, pastas e outros produtos alimentícios. Devido à alta procura desta pimenta, torna-se necessária a obtenção de sementes de qualidade a fim de se obter maior produtividade. Vários fatores podem influenciar na qualidade das sementes, de modo que a utilização de modelos de regressão logística para predizer a percentagem de germinação das sementes pode trazer resultados importantes acerca das melhores condições de produção. O presente trabalho objetivou comparar modelos de regressão logística utilizando a metodologia da curva Receiver Operating Characteristic (ROC) para avaliar a capacidade germinativa de sementes de Capsicum chinense Jacquin. Os fatores avaliados foram: cultivar (laranja e vermelho), período de armazenamento do fruto (1, 7 e 14 dias), método de extração (manual e mecânico) e período de armazenamento das sementes (3, 6, 9 e 12 meses). Pelo critério de informação de Akaike (AIC) e área abaixo da curva ROC, o melhor modelo foi aquele com interação entre “cultivar”, “armazenamento da semente” e “armazenamento do fruto”, sem o fator “método de extração”. O modelo permitiu determinar a chance de germinação a partir da combinação entre estes fatores.

RESUMO A ampliação da base genética dos materiais de cultivo é trabalho contínuo do programa de melhoramento genético do cupuaçuzeiro (Theobroma grandiflorum). Há necessidade, entretanto, que os parentais envolvidos estejam caracterizados fenotípica e genotipicamente, para auxiliar na escolha dos indivíduos que serão hibridizados, bem como para garantir os mecanismos de rastreabilidade e proteção das cultivares. Este estudo teve como objetivo identificar e selecionar clones de cupuaçuzeiro inter-compatíveis e geneticamente divergentes, utilizando como ferramentas as distâncias genotípica, fenotípica e combinada mistas, assim como as taxas de compatibilidade entre os clones. Estimativas das distâncias genéticas entre os clones foram realizadas com base em 14 iniciadores microssatélites heterólogos de cacaueiro (Theobroma cacao) que amplificam o DNA do cupuaçuzeiro. Para a caracterização fenotípica foram empregadas 14 variáveis relacionadas à produção de frutos. A matriz de dissimilaridade conjunta foi obtida por meio da soma das matrizes de dissimilaridade fenotípica e molecular. A compatibilidade intra e inter-clonal foi estimada através de cruzamentos controlados. Houve uma baixa correlação entre as matrizes de dissimilaridade com base nos dados moleculares e agronômicos. Quanto à compatibilidade, todos os clones foram auto-incompatíveis, contudo, compatíveis entre si, com diferentes taxas. O índice de compatibilidade foi fortemente influenciado pelo grau de relacionamento dos clones. Foi possível identificar e selecionar os conjuntos de clones de cupuaçuzeiro mais promissores para ser usados em melhoramento genético, apesar da ocorrência de relação genética entre eles.

ABSTRACT The expansion of the genetic base of cultivated materials is an ongoing activity of the cupuassu (Theobroma grandiflorum) breeding program. However, the parents involved need to be genotypically and phenotypically characterized to ensure compatibility of crossings, as well as to assist in the selection of more promising individuals for hybridization. This study aimed to identify and select T. grandiflorum clones that are compatible and genetically divergent using tools such as the estimates of genotypic, phenotypic, and combined distances, as well as the compatibility rates among clones. The genetic distance analysis of the clones was performed with 14 heterologous microsatellite primers of cocoa (Theobroma cacao) that amplify the DNA of cupuassu. Phenotypic characterization was based on 14 variables related to fruit production. The joint dissimilarity matrix was obtained by means of the sum of the phenotypic and molecular dissimilarity matrices. The intra- and inter-clonal compatibility was estimated through controlled crossings. A low correlation was noted between the dissimilarity matrices based on the molecular and agronomic data. As for compatibility, all clones were self-incompatible, with different compatibility rates when crossed. The compatibility index was strongly influenced by the degree of relationship of the clones. It was possible to identify and select the most promising sets of cupuassu clones to be used in breeding programs, despite their genetic relationship.

This study aimed at investigating the genetic divergence of eighteen accessions of cupuaçu trees based on fruit morphometric traits and comparing usual methods of cluster analysis with the proposed multiscale bootstrap resampling methodology. The data were obtained from an experiment conducted in Tomé-Açu city (PA, Brazil), arranged in a completely randomized design with eighteen cupuaçu accessions and 10 repetitions, from 2004 to 2011. Genetic parameters were estimated by restricted maximum likelihood/best linear unbiased prediction (REML/BLUP) methodology. The predicted breeding values were used in the study on genetic divergence through Unweighted Pair Cluster Method with Arithmetic Mean (UPGMA) hierarchical clustering and Tocher’s optimization method based on standardized Euclidean distance. Clustering consistency and optimal number of clusters in the UPGMA method were verified by the cophenetic correlation coefficient (CCC) and Mojena’s criterion, respectively, besides the multiscale bootstrap resampling technique. The use of the clustering UPGMA method in situations with and without multiscale bootstrap resulted in four and five clusters, respectively, while the Tocher’s method resulted in seven clusters. The multiscale bootstrap resampling technique proves to be efficient to assess the consistency of clustering in hierarchical methods and, consequently, the optimal number of clusters.