ONTraC characterizes spatially continuous variations of tissue microenvironment through niche trajectory analysis

Area of Science:

• Computational biology
• Bioinformatics
• Spatial transcriptomics

Background:

• Single-cell resolution technologies allow detailed mapping of tissue spatial organization.
• Existing methods lack robust analysis of spatially continuous tissue microenvironments.

Purpose of the Study:

• Introduce ONTraC, a novel graph neural network framework.
• Enable construction and analysis of spatial trajectories at the niche level.
• Characterize structural and functional organization of tissue microenvironments.

Main Methods:

• Developed ONTraC, a graph neural network-based framework.
• Utilized simulated and real biological datasets for benchmarking.
• Analyzed gene expression, regulatory networks, and cell-cell interactions.

Main Results:

• ONTraC outperforms existing methods in analyzing spatial microenvironments.
• Successfully captures normal anatomical structures and disease-associated changes.
• Identifies microenvironment-dependent shifts in molecular and cellular patterns.

Conclusions:

• ONTraC provides a powerful framework for spatial microenvironment analysis.
• Enables deeper understanding of tissue organization and function.
• Facilitates discovery in normal and diseased tissues.