Métodos Computacionais para Análise e Caracterização de Imagens de Embriões da Drosophila melanogaster

• Ano de defesa: 2020
• Autor(a) principal: Pereira, Daniela Justiniano de Sousa
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Ciência da Computação
• Palavras-chave em Português: Processamento de imagens; Image processing; Análise de Imagens; Image analysis; Núcleos celulares; Cell nuclei; Expressão gênica; Gene expression; Segmentação de imagens; Images segmentation; Microscopia confocal; Confocal microscopy; Drosophila melanogaster; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO; Computação
• Link de acesso: https://repositorio.ufu.br/handle/123456789/31136; http://doi.org/10.14393/ufu.te.2021.1

In this thesis, it is proposed a computational methodology for processing and analysis of Drosophila melanogaster embryos images. Basic investigations of the gene expression in this organism provide fundamental knowledge about important biological processes, such as cell development and differentiation. The methodology consists of six computational modules developed from Digital Image Processing techniques that can be applied sequentially or individually, namely: (1) Embryo isolation, (2) Standardization, (3) Nuclear segmentation, (4) Representation of nuclear data, (5) Representation of cytoplasm data and (6) Expression quantification. The modules include solutions for data standardization, image segmentation, representation of gene expression data and extraction of quantitative gene expression data – being useful for biological analysis ranging from the identification of the main embryo in the image to the visualization of the gene expression patterns. Together, it is offered a generic solution for the treatment of the data complexity of this image type. The proposal was validaded using embryo surface images and sagittal and transverse sections. These images were obtained from public image databases (BDGP and FlyEx) and also from our own databases. The results show that the proposed methodology is flexible and robust, as it handles a wide variety of images in this domain. Part of the proposed methods performed better than those found in the literature. In addition, it is presented the biological interpretations made from the data obtained with the application of the methodology.