Current Bioinformatics Seminar
Time: 11AM Tuesdays.
Venue: Davis Auditorium and Online
5 May 2026
AI Image-Omics for Genetic Discovery of the Human Spine and Spinal Cord
Jiru Han and Pin SunWEHI
Population-based biobanks provide many novel discovery opportunities to understand the genetic architecture of complex human traits. The UK Biobank is unique in the breadth of its available data, which includes multiple medical imaging modalities, genomics, and health-related outcomes. In this bioinformatics seminar, we present two applications leveraging automated deep learning frameworks to process three imaging modalities from the UK Biobank. In the first application, we provide proof of concept that a combination of body MRI and DXA allows for the characterization of the spinal column and rib anatomy. In the second application, we use the UK Biobank brain MRI to examine the shape metrics of the upper spinal cord to conduct the largest genome-wide association study of the spinal cord to date. Both applications benefited from AI applications to biomedical image analysis, which allowed the segmentation, annotation, and multimodal integration to facilitate discovery in large-scale imaging datasets. We will go beyond the biological discoveries to focus on what happened behind the scenes, unpacking the technical considerations, data integration strategies, and methodological choices required for large-scale Image-Omics. We hope to showcase how these pipelines were successfully built through a deeply collaborative effort blending neuroimaging and genetics expertise.