Identification of high-risk cells in single-cell spatially resolved transcriptomics data using Diagnostic Evidence...

Area of Science:

• Genomics
• Computational Biology
• Pathology

Background:

• Spatially resolved transcriptomics (SRT) provides insights into disease processes by examining tissue samples.
• Current methods can identify cell types associated with disease but struggle to link individual cells to specific disease attributes.

Purpose of the Study:

• To introduce and evaluate Diagnostic Evidence GAuge of Single-cells and spatial transcriptomics (DEGAS), a novel computational method.
• To demonstrate DEGAS's capability in associating individual cells with disease attributes across various single-cell spatially resolved transcriptomics (scSRT) platforms.

Main Methods:

• DEGAS utilizes latent representations of gene expression data and domain adaptation techniques.
• The method transfers disease attributes from patient data to individual cells within single-cell RNA sequencing (scRNA-seq) datasets.
• DEGAS was tested on liver hepatocellular carcinoma, skin cutaneous melanoma, and a new Type II Diabetes Xenium dataset.

Main Results:

• DEGAS successfully identified high-risk cells and regions in liver and skin cancer datasets, validated by known markers.
• The method was applied to a Type II Diabetes Xenium dataset, revealing high-risk cells within the tissue samples.
• DEGAS demonstrated versatility in adapting to data from different scSRT platforms.

Conclusions:

• DEGAS effectively identifies high-risk cells and regions in tissue samples, offering significant advancements in disease process analysis.
• The method overcomes limitations of existing approaches by enabling the association of individual cells with disease attributes.
• DEGAS holds promise for deeper understanding and diagnosis of various diseases through precise cellular analysis.