Signal-based spatial domain identification of spatially resolved transcriptomics with multigraph fusion

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Spatially resolved transcriptomics (SRT) offers insights into tissue micro-environments by analyzing cell transcriptomes within intact tissues.
• Current SRT domain identification methods often overlook intercellular interactions, leading to suboptimal accuracy and sensitivity.
• Accurate identification of spatial domains is crucial for understanding tissue organization and function.

Purpose of the Study:

• To develop a novel framework, SiDMGF (Signal-based Domain identification with Multi-Graph Fusion), for enhanced spatial domain identification from SRT data.
• To improve the accuracy and robustness of spatial domain identification by integrating diverse biological data.
• To validate the performance of SiDMGF against existing state-of-the-art methods.

Main Methods:

• SiDMGF integrates gene set-derived signaling graphs and spatial graphs to model biological context, spatial information, and gene expression.
• The framework utilizes multi-graph fusion to jointly analyze these different data modalities.
• Performance was evaluated on multiple benchmark SRT datasets and diverse spatial sequencing platforms.

Main Results:

• SiDMGF consistently outperformed existing state-of-the-art methods in spatial domain identification across various datasets.
• The framework demonstrated superior accuracy and robustness in identifying spatial domains.
• SiDMGF was effectively applied to cancer tissue samples, accurately delineating tumor micro-environment heterogeneity.

Conclusions:

• SiDMGF provides a significant advancement in spatial domain identification for SRT data.
• The integration of signaling and spatial graphs enhances the understanding of tissue micro-environments.
• SiDMGF shows promise for applications in cancer research and precision medicine.