Ανάλυση σημασίας σηματοδοτικών υπο-δικτύων και συσχέτιση τους με το μηχανισμό δράσης
φαρμακολογικών ουσιών

Κολέρη, Χριστίνα; Koleri, Christina

dc.contributor.author	Κολέρη, Χριστίνα	el
dc.contributor.author	Koleri, Christina	en
dc.date.accessioned	2023-09-25T10:20:49Z
dc.date.available	2023-09-25T10:20:49Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/58093
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.25790
dc.rights	Default License
dc.subject	Νευρωνικά Δίκτυα	el
dc.subject	Μηχανική Μάθηση	el
dc.subject	Συστημική Βιολογία	el
dc.subject	Machine Learning	en
dc.subject	Neural Networks	en
dc.subject	System Biology	en
dc.title	Ανάλυση σημασίας σηματοδοτικών υπο-δικτύων και συσχέτιση τους με το μηχανισμό δράσης φαρμακολογικών ουσιών	el
dc.title	Analysis of important signaling sub-networks for drug mechanism of action identification	en
dc.contributor.department	BioSysLab	el
heal.type	bachelorThesis
heal.classification	Μηχανική Μάθηση	el
heal.classification	Machine Learning	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-02-24
heal.abstract	Deep learning has emerged as a powerful tool in the era of Big Data, addressing complex challenges across various research domains. This thesis focuses on the application of deep learning techniques to unravel the mechanisms of action for selected drugs using biological signaling networks. We present a comprehensive pipeline that identifies significant sub-networks within compound-induced signaling networks. By employing an unsupervised graph deep learning pipeline called deepSNEM, we transform compound-induced signaling networks into high-dimensional representations. Utilizing the deepSNEM embeddings and clustering with the k-means algorithm, distinct clusters enriched for specific inhibitors (mTOR, topoisomerase, HDAC, and protein synthesis inhibitors) are identified. Additionally, our pipeline incorporates a subgraph importance analysis, revealing critical nodes and subgraphs directly associated with the most prevalent mechanisms of action within each cluster. This analysis provides an interpretable framework for understanding the significance of individual proteins in the pathway. To demonstrate the practical utility of our approach, we apply deepSNEM and the subgraph importance pipeline to compounds' gene expression profiles from various experimental platforms. The results indicate that accurate hypotheses can be generated regarding the mechanisms of action for these compounds. In summary, our research offers an advanced methodology that combines deep learning techniques with signaling pathway data. By analyzing important signaling sub-networks, our pipeline contributes to the identification of drug mechanisms of action, providing valuable insights into the underlying processes driving compound effects.	en
heal.advisorName	Αλεξόπουλος, Λεωνίδας	el
heal.committeeMemberName	Προβατίδης, Χριστόφορος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Μηχανολόγων Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	55 σ.	el
heal.fullTextAvailability	false
heal.fullTextAvailability	false
heal.fullTextAvailability	false