Staining-free Visualization of Cell Boundaries in Sliced Tissues for Accurate Binning of Gene Expression in Spatial Transcriptomics

Overcoming the limitations of sparse expression data in high-resolution spatial transcriptomics remains a major challenge for dimensionality reduction and clustering analysis. Visualizing or inferring tissue cell boundaries using dye staining on the same slice can help assign capture spots on the chip to corresponding unit cells and enhance UMI counts and the number of detected genes. However, the staining process may disrupt tissue structures or degrade mRNA, leading to inaccurate spot assignment and cell annotation. In this study, we employed LysM-Cre mT/mG transgenic mice for spatial transcriptomic analysis, where cell boundaries were delineated by tdTomato signals, and myeloid cells were labeled with enhanced green fluorescent protein (EGFP). With the aid of these fluorescent markers, UMI counts and the number of detected genes per cell exceeded those obtained through naive binning with a square matrix. This staining-free approach for visualizing cell boundaries and landmark proteins provides a valuable framework for the studies of single-cell spatial transcriptomics.