Trajectory analysis of single-cell RNA-seq and imaging data reveals alternate origins of tuft cells in the gut

Understanding the origin of tissue-level heterogeneity is critical for managing the heterogeneous nature of complex diseases such as cancer and inflammatory bowel disease. We present a new computational tool that maps single-cell data onto progression trajectories that potentially model the transitional processes of cell fate specification. Our algorithm, p-Creode, uses a hierarchical placement strategy to establish transitional relationships among progenitor cell states, and uses a unique ensemble approach to produce multi-branching trajectories. The robustness of each result can be statistically assessed by a scoring metric that compares graphs with both differing nodes and connections. We successfully applied p-Creode to mass cytometry, multiplex immunofluorescence (MxIF), and single-cell RNA-seq data to extract cell transition trajectories. We employed p-Creode to investigate the origin of intestinal tuft cells, a rare chemosensory cell type thought to arise from a secretory cell lineage. By extracting p-Creode trajectories from MxIF single-cell data, we found that tuft cells can be specified outside of the Atoh1-dependent secretory lineage in the small intestine, whereas, in the colon, they arise from an Atoh1-driven developmental program. These studies introduce p-Creode as a reliable method for analyzing large datasets that depict branching transition trajectories.