IAMSAM: image-based analysis of molecular signatures using the Segment Anything Model.

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Spatial transcriptomics enables gene expression analysis within tissue context.
• Understanding molecular patterns in tissue architecture is crucial for biological research.
• Existing tools may lack user-friendliness or specific focus on morphological features.

Purpose of the Study:

• To introduce IAMSAM, a web-based tool for spatial transcriptomics data analysis.
• To enable semi-automatic selection of regions of interest based on morphological features.
• To facilitate downstream analyses like differential gene expression and cell type prediction.

Main Methods:

• Utilized the Segment Anything Model for accurate tissue image segmentation.
• Developed a user-friendly web interface for data analysis.
• Integrated tools for differential gene expression, enrichment analysis, and cell type prediction.

Main Results:

• IAMSAM accurately segments tissue images.
• The tool allows semi-automatic selection of regions based on morphological signatures.
• IAMSAM successfully performs downstream analyses within selected regions.

Conclusions:

• IAMSAM provides a streamlined approach for exploring and interpreting heterogeneous tissues.
• The tool empowers researchers with a user-friendly platform for spatial transcriptomics analysis.
• IAMSAM enhances the study of molecular patterns within tissue architecture.