Related Experiment Video
Updated: Jun 13, 2025
10:12
Droplet Barcoding-Based Single Cell Transcriptomics of Adult Mammalian Tissues
Published on: January 10, 2019
18.5K
FICTURE: scalable segmentation-free analysis of submicron-resolution spatial transcriptomics
Yichen Si1, ChangHee Lee2, Yongha Hwang3,4
1Department of Biostatistics and Center for Statistical Genetics, University of Michigan, Ann Arbor, MI, USA. ycsi@umich.edu.
Nature Methods
|September 12, 2024
Summary
FICTURE is a new segmentation-free method for analyzing high-resolution spatial transcriptomics (ST) data. It efficiently reveals complex tissue architecture, overcoming limitations of existing cell segmentation approaches for whole-transcriptome analysis.
Area of Science:
- Genomics
- Bioinformatics
- Computational Biology
Background:
- Spatial transcriptomics (ST) enables gene expression analysis at submicron resolution.
- High-resolution ST analysis faces challenges with complex tissue structures and cell segmentation.
- Existing methods struggle with irregular cell shapes and lack scalability for whole-transcriptome analysis.
Purpose of the Study:
- To introduce FICTURE, a novel segmentation-free spatial factorization method for high-resolution ST data.
- To address the limitations of current cell segmentation techniques in complex tissues.
- To enable scalable, whole-transcriptome analysis compatible with diverse ST technologies.
Main Methods:
- Developed FICTURE (Factor Inference of Cartographic Transcriptome at Ultra-high REsolution), a segmentation-free spatial factorization approach.
- Utilized a multilayered Dirichlet model for stochastic variational inference of pixel-level spatial factors.
- Ensured compatibility with both sequencing-based and imaging-based ST data.
Main Results:
- FICTURE demonstrates orders of magnitude greater efficiency compared to existing methods.
- Successfully revealed microscopic ST architecture in challenging tissues (vascular, fibrotic, muscular, lipid-laden).
- Overcame limitations of previous methods in analyzing complex tissue structures.
Conclusions:
- FICTURE offers a powerful, scalable, and precise tool for exploring high-resolution ST data.
- Its cross-platform generality makes it widely applicable to various ST datasets.
- Enables deeper understanding of tissue architecture at the cellular and subcellular levels.
