Graph Attention for Alzheimer's Disease Gene Prioritization

Master thesis (2023)

Authors

T. Verlaan Electrical Engineering, Mathematics and Computer Science

Contributors

M.J.T. Reinders Pattern Recognition and Bioinformatics - EEMCS (supervisor 1)
G.A. Bouland Pattern Recognition and Bioinformatics - EEMCS (supervisor 2)
Faculty
Electrical Engineering, Mathematics and Computer Science
Copyright: © 2023 Timo Verlaan
More Info
expand_more

Published Date

10-07-2023

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Abstract

Identifying key genes in Alzheimer’s Disease (AD) is important in increasing understanding about its pathogenesis, and discovering potential therapeutic targets. Recent advances in single-cell RNA sequencing (scRNAseq) technology have provided unprecedented opportunities to study the molecular mechanisms underlying AD at the cellular level. In this study, we have trained a Graph Attention Network (GAT) to predict disease status of single cells from the SEA-AD scRNAseq dataset. Furthermore, we propose a method for interpreting the learned attention weights of the GAT to score genes on their importance for the prediction, treating gene prioritization as a feature importance problem. We have identified several genes associated with AD, including RBFOX1, NRG1, NRG3, GPC6, HNRNPC and CSMD1. We also found significant gene set enrichment in several terms related to AD and dementia, warranting future research about the presented top genes.

Files

Thesis.pdf
(.pdf | 9.87 Mb)