Cell type identification in spatial transcriptomics data can be improved by leveraging cell-type-informative paired...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Spatial transcriptomics enables detailed gene expression analysis within tissue microenvironments.
• Current spatial transcriptomics methods face limitations in distinguishing similar cell types and in low-transcript capture regions.

Purpose of the Study:

• To develop a novel computational methodology for integrating spatial transcriptomics data with paired tissue images.
• To enhance the accuracy of cell type composition inference in spatial transcriptomics data.

Main Methods:

• A statistical approach was developed to computationally integrate spatial transcriptomics data with cell-type-informative paired tissue images.
• The methodology was demonstrated using immunofluorescence markers on mouse brain tissue and AI-annotated H&E images of breast cancer tissue.

Main Results:

• The integration markedly improved the identification of clinically relevant immune cell infiltration in breast cancer.
• The approach demonstrated enhanced cell type identification in spatial transcriptomics data.

Conclusions:

• Combining spatial transcriptomics with paired tissue imaging offers a powerful strategy to improve cell type identification.
• This integrated approach has the potential to advance applications of spatial transcriptomics that depend on precise cell type mapping.