A spatial transcriptomics dataset of pancreas sections in normal glucose tolerance and type 2 diabetic donors
- Nick Howell
- Zoe Weiss
- Lori L. Bonnycastle
- Caleb M. Grenko
- Davide Randazzo
- Christopher H. Dampier
- Neelam Sinha
- Narisu Narisu
- Amy J. Swift
- Michael R. Erdos
- Leslie G. Biesecker
- Francis S. Collins
- Catherine C. Robertson
- D. Leland Taylor
2025-09-01
Understanding the spatial distribution of gene expression in the pancreas is essential for establishing the molecular basis of pancreatic function in healthy and disease contexts. Recent platforms offer a robust method for quantifying gene expression within a spatial context. Here, we report spatial transcriptomic profiling from pancreas samples obtained from three donors with type 2 diabetes (T2D) and three donors with normal glucose tolerance (NGT). Our analysis identified a major technical challenge: substantial transcript bleed of highly abundant genes (e.g., INS and GCG ) into adjacent tissue regions. We demonstrate that this bleed can be computationally corrected using probabilistic models. Our analysis highlights the importance of incorporating bleed-correction techniques in the preprocessing of spatial transcriptomic profiling data. In summary, this study provides a dataset, methods, and resources to investigate the spatial regulation of gene expression in normal and T2D-affected human pancreas.