Title: Towards Spatial Transcriptomics: A cell is known by the company it keeps

Abstract

Single cell RNA sequencing (scRNAseq) has impacted approaches for biological and medical science by providing information about heterogeneity and transition across biological processes. However, the loss of locational information by scRNAseq made it hard to explain the cause of heterogeneity and cell transition. Now, spatial transcriptomic is expected to impact more profoundly by providing additional location information of cells.

Using physically interacting cells followed by sequencing (PICseq), we show how cells influence each other during embryogenesis. We found that the transcriptome of a cell is influenced by its neighboring cell type. Based on it, we show how to predict the neighbors of a cell from scRNAseq. We also show how to handle high-resolution spatial transcriptomics data to identify previously unknown tissue architectures and genes differentially expressed in a tissue-morphology dependent manner. Expanding our idea to high-resolution spatial transcriptomics data, we show that cells change their transcriptome depending on the neighbors and the region they belong.