Related Experiment Video
Updated: Apr 18, 2026
10:16
Mining Spatial Transcriptomics Datasets using DeepSpaceDB
Published on: September 5, 2025
1.1K
A zero-inflated hierarchical generalized transformation model to address non-normality in spatially-informed
Hunter J Melton1, Jonathan R Bradley2, Chong Wu3,4
1Department of Biomedical Data Science, Geisel School of Medicine at Dartmouth, Hanover, NH 03756, United States.
Biometrics
|April 17, 2026
Summary
This study introduces a new method, ZI-HGT + CARD, to accurately identify cell types in oral cancer's tumor microenvironment using spatial transcriptomics. The approach improves analysis of zero-inflated data, crucial for understanding cancer growth.
Area of Science:
- Genomics
- Cancer Research
- Bioinformatics
Background:
- Oral squamous cell carcinomas (OSCC) present diagnostic challenges with low survival rates.
- Spatial transcriptomics is key to understanding OSCC tumor microenvironments.
- Existing cell-type deconvolution methods struggle with zero-inflated OSCC data.
Purpose of the Study:
- To develop a novel method for accurate cell-type deconvolution in zero-inflated OSCC spatial transcriptomics data.
- To improve the understanding of tumor microenvironment composition in OSCC.
- To quantify uncertainty in cell-type proportion estimations.
Main Methods:
- Developed a zero-inflated hierarchical generalized transformation model (ZI-HGT).
- Applied ZI-HGT as a Bayesian auxiliary technique to Conditional AutoRegressive Deconvolution (CARD).
- Validated the ZI-HGT + CARD framework using simulations and OSCC data analysis.
Main Results:
- The ZI-HGT + CARD framework significantly enhances cell-type deconvolution accuracy.
- The method effectively handles high zero-inflation in OSCC spatial transcriptomics data.
- Accurate quantification of cell-type proportions and their uncertainties was achieved.
- Identified locations of diverse fibroblast populations within the OSCC tumor microenvironment.
Conclusions:
- The ZI-HGT + CARD framework offers a robust solution for cell-type deconvolution in challenging OSCC spatial transcriptomics data.
- This approach provides critical insights into tumor microenvironment heterogeneity.
- Understanding fibroblast populations is vital for OSCC progression and immunosuppression research.
