Institutional Repository
Technical University of Crete
EN  |  EL

Search

Browse

My Space

A 21-hub-gene signature in multiple sclerosis identified using machine learning techniques

Tsakaneli Stavroula, Bei Aikaterini, Zervakis Michail

Full record


URI: http://purl.tuc.gr/dl/dias/1DCC5A2B-CCD5-4B80-8E61-12EF753DDEA1
Year 2022
Type of Item Conference Full Paper
License
Details
Bibliographic Citation S. Tsakaneli, E. S. Bei and M. E. Zervakis, "A 21-hub-gene signature in multiple sclerosis identified using machine learning techniques," in Proceedings of the 2022 IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI 2022), Ioannina, Greece, 2022, doi: 10.1109/BHI56158.2022.9926949. https://doi.org/10.1109/BHI56158.2022.9926949
Appears in Collections

Summary

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease that affects approximately 2.8 million persons globally. While there is currently no cure for this neurodegenerative disease, MS has become a highly manageable disease through treatment options like disease-modifying medications, that can help to control the symptoms and slow disease progression. Among them, interferon beta (IFNβ) therapy is a first-line treatment for MS but has shown to be only partially effective. Thus, it is important to identify biomarkers that aid in early identification of IFNβ responders. In this study, based on gene expression profiles from untreated and interferon treated patients from a publicly available dataset, we performed differential expression analysis and Pigengene network association (weighted correlation network analysis (WGCNA) and Bayesian networks modeling) in order to construct a high-confidence protein-protein (PPI) interaction network. Subsequently, aiming to find the most significant clustering modules and hub genes, we applied a number of topological analysis methods (cytoHubba plugin) followed by MCODE clustering algorithm. Our approach resulted in highly connected hub genes generating a reliable 21-hubgene signature that could predict the response of interferon beta (IFNβ) therapy in patients with MS. The 21-hub-gene signature showed high classification performance (Accuracy = 91,49%, Sensitivity = 94.55%, Specificity = 87.15%) demonstrating potential clinical benefit.

Services

Statistics