Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/134426
Title: | Finding a predictive gene signature in pancreatic cancer using gene expression |
Author: | Torre Pernas, Sabela de la |
Tutor: | BARCELÓ, PhD, CARLES |
Others: | Prados Carrasco, Ferran |
Abstract: | PDAC is one of the most aggressive human cancers with a 5-year overall survival rate lower than 10%. At the time of diagnose, it is already too late for many patients who can't benefit from surgical procedure or chemotherapy. The objective of the present work is to study the gene expression profiles from different perspectives, and generate several signatures with predictive power. This could allow physicians to improve prognosis and find better treatments according to each patient. The biological processes where the genes in these signatures participate were studied, showing that most of the genes are related to cellular and metabolic processes. The survival analysis with Kaplan-Meier Plotter showed that many of these genes had a significant correlation with survival in pancreatic cancer. Finally, the predictive power of the different signatures was assessed using a Machine Learning algorithm. Specifically, several Random Forest models were trained and evaluated with different configurations. The best accuracy (62%) was obtained with the common signature, which included the intersection of the genes in the signatures of each of the groups studied, Treatment vs outcome and Gene expression vs outcome. |
Keywords: | Pancreatic cancer gene expression predictive model |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | 6-Jun-2021 |
Publication license: | http://creativecommons.org/licenses/by-nc-nd/3.0/es/ http://creativecommons.org/licenses/by-nc-nd/3.0/es/ |
Appears in Collections: | Bachelor thesis, research projects, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
sdelatorrepernasTFM0621memory.pdf | Memory of TFM | 1,92 MB | Adobe PDF | View/Open |
Share:
This item is licensed under a Creative Commons License