Anàlisi del patró de disbiòsi intestinal en colitis ulcerosa a partir de mostres fecals

The work is based on metagenomic sequencing data of 16S rRNA from fecal samples of patients with ulcerative colitis (UC) and healthy controls (HC) from different previously published studies. The different studies have been divided into two cohorts: a discovery cohort (CD) and a validation cohort (C...

Descripción completa

Detalles Bibliográficos
Autor: Castany Roma, Miquel
Tipo de recurso: tesis de maestría
Fecha de publicación:2023
País:España
Institución:Universitat Oberta de Catalunya (UOC)
Repositorio:O2, repositorio institucional de la UOC
OAI Identifier:oai:openaccess.uoc.edu:10609/148589
Acceso en línea:http://hdl.handle.net/10609/148589
Access Level:acceso abierto
Palabra clave:16S rRNA
metagenomics
gut microbiome
metagenòmica
microbioma intestinal
Ulcerative colitis -- TFM
Colitis ulcerosa -- TFM
Descripción
Sumario:The work is based on metagenomic sequencing data of 16S rRNA from fecal samples of patients with ulcerative colitis (UC) and healthy controls (HC) from different previously published studies. The different studies have been divided into two cohorts: a discovery cohort (CD) and a validation cohort (CV). The relevant taxonomic classifications of each sample have been performed using specialized software for these tasks. In our case, we have used the software known as Mothur (Schloss et al., 2009) and a taxonomic classification database for 16S rRNA sequences such as SILVA (Quast et al., 2013). Once the taxonomic classification data has been obtained, statistical analysis has been carried out using Rstudio to determine and extract a common dysbiosis pattern to the samples of patients composing the discovery cohort. Subsequently, these results have been attempted to be extrapolated to the validation cohort, which consisted of UC patient samples from another study. Afterwards, a task dedicated to the prediction of clinical outcomes has been carried out using the relative abundances of different taxa as differentiating features. Machine learning techniques (ML) (Serrano-Gómez et al., 2021) have been used for this purpose, but other approaches based on generalized linear models (GLMs) or multiple regressions have also been employed (Rivera-Pinto et al., 2018).