Unit:
Department of Informatics and TelecommunicationsΠληροφορική
Author:
KANELLAKI MARIA-ANNA
Supervisors info:
ΣΤΑΜΑΤΟΠΟΥΛΟΣ ΠΑΝΑΓΙΩΤΗΣ, Επίκουρος Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικόν και Καποδιστριακόν Πανεπιστήμιον Αθηνών
Original Title:
Convolutional Neural Networks and their Application in Cancer Diagnosis based on RNA-Sequencing
Translated title:
Convolutional Neural Networks and their Application in Cancer Diagnosis based on RNA-Sequencing
Summary:
Gene expression analysis is the study of the way genes are transcribed to synthesize
functional gene products, functional RNA species, or protein products. Its study can
provide insights of cellular processes, such as cellular differentiation and abnormal
pathological processes.
Cancer is a genetic disease where genetic variations cause abnormally functioning
genes that appear to alter expression. Proteins, being the final products of gene
expression, define the phenotypes and biological processes. Therefore, detecting gene
expression levels can be used for cancer diagnosis, prognosis, and even treatment
prediction.
This thesis will be analyzing the theory and applications of Deep Learning. It will then
apply Deep Learning (DL) and in particular a Convolutional Neural Network (CNN) as a
means for the diagnosis of multiple cancer types (pan-cancer classification) using gene
expression data and specifically RNA-sequencing.
The Cancer Genome Atlas (TCGA) data, which consists of RNA-sequencing, will be
preprocessed and then embedded into multiple two-dimensional (2D) images. These
images will then be applied to a Convolutional Neural Network which will classify them
into 33 types of cancer, in an attempt to achieve the highest possible diagnosis
accuracy.
Main subject category:
Technology - Computer science
Keywords:
Deep Learning, Convolutional Neural Network, Classification, Cancer Diagnosis, Gene Expression, RNA-Sequencing
Number of references:
121
