Please use this identifier to cite or link to this item:
Title: Machine learning para la selección de genes implicados en el desarrollo de Arabidopsis thaliana utilizando datos de expresión génica
Author: Saura Sánchez, Maria Teresa
Director: Sánchez Pla, Alexandre
Tutor: Vegas Lozano, Esteban
Keywords: machine learning
Arabidopsis thaliana
Issue Date: Jan-2019
Publisher: Universitat Oberta de Catalunya (UOC)
Abstract: Transcriptional programs are important in the development of the structures throughout the life cycle of plants. DNA microarray technology has provided a useful tool to discover relevant genes in the development of the reference plant Arabidopsis thaliana. However, previous studies use a reduced number samples to discover marker genes based on its specific expression along the tissues. In this work, a machine learning approach is presented to select relevant genes in the development of A. Thaliana. A database was built with more than 500 expression profiles corresponding to seeds, seedlings, roots, leaves and flowers. Gene selection was carried out with three different ML methods: FP-RF, RF-RFE, SVM-RFE. Furthermore, an autoencoder architecture was evaluated for dimensionality reduction of the data. The genes selected by ML techniques yield high classification performance in SVM, RF and ANN algorithms. Moreover, these genes are biologically relevant to plant development process. This work provides a new approach to study plant development from gene expression data.
Language: Spanish
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
saurasanchezTFM0119memoria.pdfMemoria del TFM5.7 MBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons