Title: Building multiscale models from single-cell data to study stem cell lineage dynamics
Abstract: The rapid expansion of data gathered for individual cells (e.g. via scRNA-seq) offers great potential to characterize cell types and their interactions in new depths. Significant remaining challenges for the analysis of such data include inference of the lineage structure of — and transition probabilities between — cell states, and communication between cells. Greater insight could be gained from these data given appropriate methods to compare datasets, or to explore the dynamical systems properties of cell states. We present two methods for the analysis of scRNA-seq data to address these challenges, and demonstrate their effectiveness on hematopoietic and epithelial stem cell systems. We also develop methods for the construction of single-cell communication networks via paracrine signaling factors. Using simple models of transcriptional networks, we explore the stability properties of cell states, and demonstrate that such models can help us to extrapolate from single-cell data to biological function.