Integração de dados morfoagronômicos, moleculares e fitopatológicos para estabelecimento de coleção nuclear

• Ano de defesa: 2011
• Autor(a) principal: Martins, Francielle Alline
• Orientador(a): Carneiro, Pedro Crescêncio Souza
• Banca de defesa: Carneiro, Antônio Policarpo Souza , Bhering, Leonardo Lopes , Silva, Felipe Lopes da
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Doutorado em Genética e Melhoramento
• Departamento: Genética animal; Genética molecular e de microrganismos; Genética quantitativa; Genética vegetal; Me
• País: BR
• Palavras-chave em Português: Codificação de dados; Tomateiro; Diversidade genética
• Palavras-chave em Inglês: Data coding; Tomato; Genetic diversity
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::AGRONOMIA::FITOTECNIA::MELHORAMENTO VEGETAL
• Link de acesso: http://locus.ufv.br/handle/123456789/1322

The objective of this study was to perform the data integration of quantitative traits, multicategorical molecular phytopathological order to evaluate the genetic diversity of subsamples of tomato of Germplasm Bank of Vegetable of the Federal University of Viçosa (BGH-UFV). A data set of 67 subsamples of tomato BGH-UFV was used. It was characterized according to 27 quantitative traits, 34 qualitative characters, 144 ISSR loci, reaction to Alternaria solani, Pseudomonas syringae pv. tomato and Tomato yellow spot virus (ToYSV). The assessment of genetic diversity among the subsamples was performed for each set of characters by Tocher grouping and then proceeded to the comparison between the dissimilarity matrices obtained from the data of different natures by the Mantel Z test. Although the correlation values between the matrices have been significant at 5% probability, these were of low magnitude, offering no support for extrapolating results from one dataset to another. Thus, two strategies for the integration of data were used: CONV - conversion of quantitative and phytopathologi al data in multicategorical order to obtain a single dissimilarity matrix that encompassed all the characters regardless of their nature and SOMA - obtain the dissimilarity matrices for individual each character set and the sum of them. Seventeen subxii strategies for converting data were analyzed; DEA-3 (fair division of amplitude in three classes) stands out among these, whose correlation value with the original dissimilarity matrix was 0.782. By comparing the strategies SOMA and CONV, it was observed high correlation coefficient, r = 0.96, between the dissimilarity matrices obtained by each. However, the strategy CONV stood out, since it allowed greater discrimination of the subsamples. Were established 10 subcollections from 67 subsamples of tomato from BGH-UFV. These subcollections were defined by combining the nature of data collected and sampling intensity. The sampling strategy used was the logarithm stratified, and the subsamples selected for each stratum were determined by multivariate analysis from the grouping method of Tocher reversed. The intensities were evaluated from 20 to 30%. To assess the representativeness of subcollections set and setting that best represented the subsamples of tomato BGH-UFV comparisons were made between the initial collection and subcollections, considering the variance and the index of coincidence of the amplitude. The graphical analysis of variability was performed by calculating the frequency of subsamples represented in each class of quantitative and multicategoric traits, previously converted into binary, and molecular characteristics. Among subcollections at 20% intensity, it stood out the subcollection based data integration CONV-20, for own the highest rates of coincidence of the amplitude followed by values of variance more appropriate. The 30% intensity, the subcollection MOL-30 was as efficient as the subcollections based on integration of data, when only the index of coincidence of the amplitude and the average of the variances were considered. However, the graphical analysis of variability showed a slight superiority of the subcollection CONV-30 in maintaining the variability, especially regarding the characters multicategoric. Where data from different sources are available, the establishment of core collections based on the integration of these data should be prioritized.