| Title: | Harmonizing Multi-Modal Biological Data through Manifold Integration and Deep Learning for Enhanced Prediction of Diseases |
| Lead Institution: | National University of Singapore |
| Principal investigator: |
Application 146760
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
About
Our research aims to transform the way we predict complex neurological diseases like Alzheimer's by leveraging the rich, multi-modal biological data available in the UK Biobank. This data is inherently complex and exists in high-dimensional spaces, making it challenging to interpret. However, our research posits that these diverse datasets can be harmonized into a singular, low-dimensional representation, known as a manifold, which captures the essential features of the data while sidelining noise and redundancies. We plan to use manifold learning techniques to transition this high-dimensional data into a more digestible, lower-dimensional space. This manifold will integrate multiple types of data-ranging from genetic sequences to imaging scans-thereby offering a more holistic view of disease mechanisms. Once this unified manifold is created, we will employ deep learning algorithms, specifically tailored neural network architectures, to analyze it. Deep learning is particularly adept at discerning intricate patterns in extensive datasets, making it an ideal tool for this manifold-integrated data. The fusion of manifold learning and deep learning in our approach promises groundbreaking advancements in disease prediction and diagnosis. By focusing on multi-modal data and manifold integration, we aim to significantly amplify the precision and effectiveness of disease prediction models. This research project is slated for a duration of 3 years. The end goal is to develop a computational tool based on this integrated approach, which will be a significant step forward in the field of personalized medicine. For the public, this research offers the promise of earlier and more accurate diagnoses, enabling timely interventions and personalized treatment plans. This could lead to improved patient outcomes, reduced healthcare costs, and a better understanding of complex neurological diseases. Overall, our project aligns closely with the public interest by aiming to make healthcare more effective and accessible.1 Publication
| Pub ID | Title | Author(s) | Year | Journal |
|---|---|---|---|---|
| 15200 | Manifold fitting reveals metabolomic heterogeneity and disease associations in UK Biobank populations | Bingjie Li (+4) | 2025 | Proceedings of the National Academy of Sciences of the United States of America |
Enabling scientific discoveries that improve human health