A carregar...
Publicação
FTIR-ATR analysis of covid-19 blood plasma for immunological biochemical indicators estimation
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.8 MB | Adobe PDF |
Orientador(es)
Resumo(s)
Neste estudo, avaliamos a exatidão dos resultados, ao usar a Espectroscopia de Infravermelho com Transformada de Fourier com Reflexão Total Atenuada, juntamente com análise multivariada, para analisar e identificar amostras de plasma de pacientes positivos e negativos para a CoVid-19.
O processo de triagem de CoVid-19 tornou-se um processo difícil, devido ao aumento de casos e, consequentemente, a afluência aos hospitais. Com isto, a hipótese de utilizar esta técnica como ferramenta diagnóstica, poderia ser um método mais rápido para identificar pacientes com indicadores inflamatórios mais elevados.
Primeiramente, foi realizada uma análise exploratória utilizando a Análise de Componentes Principais, utilizando os espectros de amostras de pacientes de diversas origens, vacinados com as vacinas Pfizer/BioNTech e AstraZeneca.
Numa fase seguinte, foi desenvolvida uma Análise Discriminante de Mínimos Quadrados Parciais, dividida em seis modelos (três sem restrição de comprimentos de onda e três em que os comprimentos de onda foram restritos a 1670-900 cm-1). Nesta análise utilizámos as mesmas amostras que na Análise de Componentes Principais, e o método foi capaz de diferenciar as amostras negativas para a CoVid-19 das amostras positivas, com uma média de exatidão de 85%. Foi possível concluir que as zonas correspondentes ao comprimento de onda referentes a proteínas e carboidratos foram as principais características consideradas para classificar as amostras.
Os resultados sugerem que é possível distinguir entre amostras positivas e negativas, utilizando tanto a Análise de Componentes Principais como Análise Discriminante de Mínimos Quadrados Parciais, o que sustenta a hipótese de que a Espectroscopia de Infravermelho com Transformada de Fourier com Reflexão Total Atenuada pode ser uma metodologia promissora para o diagnóstico da doença de CoVid-19.
In this study, we evaluated the accuracy of outcomes, when using Fourier Transform Infrared with Attenuated Total Reflection spectroscopy, coupled with multivariate analysis, to analyze and identify plasma samples from patients with Coronavirus Disease 2019 positive PCR test and no reported positive PCR test. Triage process of Coronavirus Disease 2019 has become a difficult process, since the increasing of cases and, consequently, the affluence to the hospitals. With this, the hypothesis of using this technic as a diagnostic tool could be a faster method to identify patients with higher inflammatory indicators. Firstly, an exploratory analysis using Principal Component Analysis was performed utilizing the spectra of samples from diverse sources patients, vaccinated with Pfizer/BioNTech and AstraZeneca vaccines. In a next phase, a Partial Least Squares-Discriminant Analysis was developed, divided into six models (three with no wavenumber restriction and three in which wavenumber were restricted to 1670-900 cm -1). We employed the same samples as in Principal Component Analysis, and the method was able to differentiate the Coronavirus Disease 2019 negative samples from Coronavirus Disease 2019 positive samples, with an accuracy media of 85%. It was possible to conclude that proteins and carbohydrates wavelength correspondent zones were the main features considered to classify the samples. The results suggest that it is possible to distinguish between positive and negative Coronavirus samples, using Principal Components Analysis and Partial Least Square-Discriminant Analysis, which supports the hypothesis that Transform Infrared with Attenuated Total Reflection spectroscopy can be a promising methodology for Coronavirus Disease 2019 diagnosis.
In this study, we evaluated the accuracy of outcomes, when using Fourier Transform Infrared with Attenuated Total Reflection spectroscopy, coupled with multivariate analysis, to analyze and identify plasma samples from patients with Coronavirus Disease 2019 positive PCR test and no reported positive PCR test. Triage process of Coronavirus Disease 2019 has become a difficult process, since the increasing of cases and, consequently, the affluence to the hospitals. With this, the hypothesis of using this technic as a diagnostic tool could be a faster method to identify patients with higher inflammatory indicators. Firstly, an exploratory analysis using Principal Component Analysis was performed utilizing the spectra of samples from diverse sources patients, vaccinated with Pfizer/BioNTech and AstraZeneca vaccines. In a next phase, a Partial Least Squares-Discriminant Analysis was developed, divided into six models (three with no wavenumber restriction and three in which wavenumber were restricted to 1670-900 cm -1). We employed the same samples as in Principal Component Analysis, and the method was able to differentiate the Coronavirus Disease 2019 negative samples from Coronavirus Disease 2019 positive samples, with an accuracy media of 85%. It was possible to conclude that proteins and carbohydrates wavelength correspondent zones were the main features considered to classify the samples. The results suggest that it is possible to distinguish between positive and negative Coronavirus samples, using Principal Components Analysis and Partial Least Square-Discriminant Analysis, which supports the hypothesis that Transform Infrared with Attenuated Total Reflection spectroscopy can be a promising methodology for Coronavirus Disease 2019 diagnosis.
Descrição
Trabalho Final de Mestrado Integrado, Ciências Farmacêuticas, 2022, Universidade de Lisboa, Faculdade de Farmácia.
Palavras-chave
Covid-19 Vaccines Infrared spectroscopy Biofluids Plasma Mestrado integrado - 2022