Mineração de dados em triagem neonatal de hemoglobinopatias
Carregando...
Arquivos
Data
2006-02-21
Autores
Melo, Luciane Moreno Storti de [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Nos anos de 1997 a 2002 foi desenvolvido um programa de triagem de hemoglobinopatias pelo Laboratório de Hemoglobinas e Genética das Doenças Hematológicas - LHGDH, UNESP, com 5976 neonatos nascidos no Hospital de Base de São José do Rio Preto. O montante de dados gerados em todas as fases dos seis anos de programa constitui uma fonte de informações, que armazenadas em banco de dados, permitem a análise por métodos computacionais na tentativa de estabelecer comportamentos, tendências e padrões, objetivando dessa forma, contribuir com a Triagem Neonatal no país, fornecendo bases para agilizar os diagnósticos e torná-los cada vez mais precisos. Após desenvolvimento de uma Base de Dados Eletrônica e inserção dos resultados da triagem de hemoglobinopatias objetivou-se: avaliar por meio de ferramenta data mining- CLIBIA as metodologias utilizadas, testar a aplicabilidade do banco de dados e analisar os dados quantitativos de hemoglobinas obtidos por HPLC nos diferentes fenótipos, em sistema de visualização 3D - Fas MapDB. A utilização de um banco de dados em laboratórios com grande fluxo de amostras mostrou-se imprescindível para a segurança, confiabilidade e agilidade das análises e ferramentas que promovem a mineração dos dados, como a CLIBIA, revelaram a necessidade de reavaliar as metodologias de rotina utilizadas no diagnóstico de hemoglobinopatias em neonatos. As visualizações tridimensionais revelaram a existência de dois padrões de distribuição das amostras com Hb normais, devido a amplitude de variação nos valores obtidos por HPLC para o pico de F1. Permitiu a separação das amostras com Hb normais e com alfa talassemia, ressaltando as diferenças estatísticas nos valores percentuais médios dos picos de Fast e A, reforçando a necessidade de melhor avaliação dos cromatogramas obtidos e a possibilidade de aplicação dessa metodologia como parte do diagnóstico de alfa talassemia em neonatos.
Since 1997 until 2002, a neonatal screening program of hemoglobinopatias were carried out in the Hemoglobin and Hematologic Genetic Diseases Laboratory - LHGDH, UNESP, with 5976 newborns from Hospital de Base de São José do Rio Preto. The amount of data obtained during six years of program constitutes a source of information, that stored in data base, allows the analysis for computational methods in the attempt to establish behaviors, trends and standards, aiming to contribute with the Neonatal screening in the country and supplied bases to better diagnostic and to become them each time more exact. After an Electronic Database were development and the screening results were inserted was aimed: to evaluate by data mining tool - CLIBIA the methodologies used, test the applicability of a hemoglobinopathies database and to analyze the quantitative hemoglobins data obtained by HPLC in the different phenotypes, in 3D visualization system - FastMapDB, for to analyze the phenotypes behavior. The use of a database in laboratories with great sample routine were essential for the security, trustworthiness and agility of the analyses and tools that promote the datamining, as the CLIBIA, had disclosed the necessity to reevaluate the methodologies used in hemoglobinopathies diagnosis routine in newborns. The three-dimensional visualizations carried out by the FastMapDB tool showed the existence of two distribution standards in the samples with normal Hb, due to the variation amplitude into the values obtained by HPLC for the F1 window. It allowed separating the samples with normal Hb from alpha thalassemia, evidencing the statistical differences in the mean values of the Fast and A window, emphasizing the necessity of better chromatograms evaluation and the possibility of this methodology application as part of alpha thalassemia diagnosis in newborns.
Since 1997 until 2002, a neonatal screening program of hemoglobinopatias were carried out in the Hemoglobin and Hematologic Genetic Diseases Laboratory - LHGDH, UNESP, with 5976 newborns from Hospital de Base de São José do Rio Preto. The amount of data obtained during six years of program constitutes a source of information, that stored in data base, allows the analysis for computational methods in the attempt to establish behaviors, trends and standards, aiming to contribute with the Neonatal screening in the country and supplied bases to better diagnostic and to become them each time more exact. After an Electronic Database were development and the screening results were inserted was aimed: to evaluate by data mining tool - CLIBIA the methodologies used, test the applicability of a hemoglobinopathies database and to analyze the quantitative hemoglobins data obtained by HPLC in the different phenotypes, in 3D visualization system - FastMapDB, for to analyze the phenotypes behavior. The use of a database in laboratories with great sample routine were essential for the security, trustworthiness and agility of the analyses and tools that promote the datamining, as the CLIBIA, had disclosed the necessity to reevaluate the methodologies used in hemoglobinopathies diagnosis routine in newborns. The three-dimensional visualizations carried out by the FastMapDB tool showed the existence of two distribution standards in the samples with normal Hb, due to the variation amplitude into the values obtained by HPLC for the F1 window. It allowed separating the samples with normal Hb from alpha thalassemia, evidencing the statistical differences in the mean values of the Fast and A window, emphasizing the necessity of better chromatograms evaluation and the possibility of this methodology application as part of alpha thalassemia diagnosis in newborns.
Descrição
Palavras-chave
Hemoglobinopatia, Neonatologia, Mineração de dados (Computação), Triagem neonatal, Data mining, Hemoglobinopatias
Como citar
MELO, Luciane Moreno Storti de. Mineração de dados em triagem neonatal de hemoglobinopatias. 2006. 88 f. Dissertação (mestrado) - Universidade Estadual Paulista, Instituto de Biociências, Letras e Ciências Exatas, 2006.