Topological velocity inference from spatial transcriptomic data

Cell fate transition is a spatiotemporal process, however, previous work has largely neglected the spatial dimension. Incorporating space and time into models of cell fate transition would be a key step toward characterizing how interactions among neighboring cells, local niche factors, and cell migration contribute to tissue development. Here, we developed topological velocity inference (TopoVelo), a computational tool to infer spatial and temporal dynamics of cell fate transition from spatial transcriptomic data. We show that TopoVelo significantly improves the accuracy and spatial coherence of inferred cell ordering compared to previous methods. TopoVelo also reveals spatial cell state dependencies of ligand-receptor genes, spatial signatures of mouse neural tubes, and patterns of early differentiation in 3D cell culture. Embryoid bodies (EBs) differentiation was performed using induced pluripotent stem cells (WTC11) using the EB formation media. After 21 days of seeding, EBs were embedded in OCT compound. 10-micron EB sections were placed on 3 mm x 3 mm CurioSeeker (Slide-seq) tiles for processing.