Alzheimer’s Disease (AD) remains a formidable challenge in neurodegenerative research, necessitating innovative approaches to uncover novel therapeutic strategies. This study presents a graph-based approach to integrate large-scale drug and protein data, aiming to identify potential drug repurposing candidates for AD. Our methodology constructs a comprehensive graph incorporating protein-protein interactions and drug-protein relations, providing a multifaceted view of the intricate relationships within the biological and pharmacological landscape. By leveraging this graph, we conduct an in-depth analysis to explore various drug repurposing possibilities, focusing on the alignment of AD-related single-cell transcriptomic data. Our approach enables the identification of promising drug candidates by examining the connectivity and interaction patterns within the graph, revealing potential therapeutic targets and drug synergies that may be beneficial for AD treatment. The integration of diverse data types allows for a more holistic understanding of the underlying molecular mechanisms and drug interactions. Through this graph-based analysis, we uncover several promising drug repurposing candidates, providing a foundation for further experimental validation and clinical investigation. This study underscores the potential of leveraging large-scale data and graph-based methodologies in drug repurposing efforts for neurodegenerative diseases, contributing to the advancement of therapeutic research in AD. Our findings illuminate the importance of integrative data analysis in biomedical research, paving the way for the development of more effective and targeted therapeutic interventions for AD.

In recent years, feature selection has become an increasingly active field of data science and machine learning research. Most of the datasets that are being used nowadays for various machine learning tasks consist of thousands of features (columns), which make them extremely complex and difficult to work with. In this paper, we propose a feature selection methodological pipeline that can be used to reduce the complexity of high dimensional datasets through the elimination of redundant and/or non-informative features as well as to improve the performance of machine learning models which are trained on high dimensional datasets. The proposed method has been applied to high-dimensional biomedical data and compared against a classic filter-based feature selection algorithm. Specifically, the method was applied to gene expression profiles of a single-cell RNA-seq dataset from healthy and infected by covid-19 human samples.