What is it about?
Spatial transcriptomics measures gene activity while preserving where cells are located in a tissue, helping researchers understand how tissues are organized. However, noisy measurements and unclear boundaries can make it difficult to accurately identify distinct tissue regions. We developed GatorTrio, a method that combines complementary information from gene expression and spatial location to better recognize these regions and reduce errors near their boundaries. Across a wide range of spatial transcriptomics datasets, GatorTrio identified tissue regions more accurately and produced clearer, more biologically meaningful spatial patterns than existing methods. This approach may help researchers better study tissue structure, cellular organization, and disease-related changes.
Featured Image
Why is it important?
Understanding where one tissue region ends and another begins is important for studying tissue organization and biological function. Yet spatial transcriptomics data can be noisy, making these boundaries difficult to identify accurately. GatorTrio offers a more flexible way to combine gene expression and spatial information, allowing the method to rely on the most useful signals in different parts of a tissue while reducing misleading connections across region boundaries. Across diverse datasets, it produced more accurate and coherent tissue maps than existing approaches. This can provide researchers with a clearer view of tissue structure and biologically meaningful spatial patterns.
Perspectives
For me, one of the most interesting aspects of this work is that spatial transcriptomics challenges us to think beyond a single definition of what makes two cells or tissue locations similar. Cells can be close in space but biologically different, while similar molecular programs can appear in different locations. This motivated us to design GatorTrio to consider several complementary relationships instead of relying on one fixed view of the tissue. I hope this work contributes to computational methods that better reflect the complexity of biological systems and makes spatial transcriptomics easier to translate into meaningful biological insights.
Yuxi Liu
Indiana University School of Medicine
Read the Original
This page is a summary of: GatorTrio: Topology-Refined Tri-View Graph Learning for Spatial Domain Identification in Spatial Transcriptomics, August 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3770855.3818918.
You can read the full text:
Contributors
The following have contributed to this page
