A thorough understanding of cellular development is incumbent on assessing the complexities of fate and kinetics of individual clones within a population. Recent clonal fate and single cell RNA-sequencing studies demonstrate that significant lineage imprinting is already in place within individual hematopoietic stem and progenitor cells (HSPCs). Dendritic cells (DCs) represent one such distinct branch of hematopoiesis and are responsible for pathogen sensing and activation of the adaptive immune response. At the population level, all three major DC subtypes including plasmacytoid DCs (pDCs), conventional DC type 1 (cDC1s) and cDC type 2 (cDC2s) can be generated from a restricted common DC progenitor (CDP) population downstream of HSPCs. However, recent clonal evidence has suggested earlier subtype-specific imprinting within the CDP and even early HSPC populations. One caveat of most of those lineage tracing studies was that clonal fate was only measured at a single time point. Therefore, questions remain as to whether the fate bias observed at one snapshot in time is consistent with earlier or later times.

Here, we develop a system for robust periodical assessment of lineage outputs of thousands of transient clones and establishment of bona fide cellular trajectories. We appraise the development of DCs in vitro from barcode-labeled HSPCs by serially measuring barcode signatures, and visualize this multidimensional data using novel developmental interpolated t-distributed stochastic neighborhood embedding (Di-SNE) time-lapse movies. We identify multiple classes of cellular trajectories of DC development that are characterized by distinct fate bias and expansion kinetics. Importantly, using clone-splitting experiments, we demonstrate that many of these cellular trajectories are ‘programed’ within individual HSPCs. Furthermore, our results demonstrate that conventional DC and plasmacytoid DC trajectories are largely separated already at the HSPC stage. This framework allows systematic evaluation of clonal dynamics and can be applied to other steady-state or perturbed developmental systems.