Studies into the ontogeny and lineage tracing of immune cells have commonly been performed in murine models using approaches such as transgenic mice, expressing reporter genes, or cell type specific genetic barcoding methods. While murine studies have offered unparalleled insights into the mechanisms of immunology, it is challenging to translate these findings to the human immune system. In addition, due to the natural constraints of human studies, the type of experiments available to investigate cell lineages in vivo are limited. Thus, most lineage studies of the human immune system are conducted either ex vivo or using in vitro culture systems. This work aimed to address this gap of knowledge by combining experimental in vivo data from human studies with mathematical models in a computational inference framework allowing us to explore a wide range of hypotheses in parallel. This work focused on two central populations in the immune system: dendritic cells (DCs) and memory T cells (TMs). We investigated the cell dynamics and potential lineage hierarchies of DCs using stable isotope labelling data. Using this data we identified essential lineage features, however uncertainty with regard to the site of pre dendritic cell (preDC), conventional dendritic cell 1 (cDC1) and conventional dendritic cell 2 (cDC2) differentiation remains. In addition, we robustly identified kinetic parameters such as the proliferation rate and average lifetime. Interestingly, preDC (≈ 1.7 days), cDC2 (≈ 1.8 days), plasmacytoid dendritic cell (pDC) (≈ 1.9 days) and cDC1 (≈ 0.5 days) in particular exhibit
relative short lifetimes in the circulation. The lineage structure of memory T cells is subject to active debate and contrasting hypotheses are suggested in the literature. The reported findings appear often biased due to the a priori hypothesis put forward by the respective authors. To address this, we developed a lineage inference framework and analysis procedure that aims to attenuate this potential source of bias while integrating different data types from various sources into a single framework. An extensive search across a wide range of different memory T cell lineage topology architectures was performed. We identified both simple as well as complex lineage features which were significantly enriched amongst the topologies most predictive of the experimental data. A simple lineage feature which was found to be conserved was that activated na ̈ıve T cells (TNs) give rise to effector memory T cell (TEM) first before they differentiate into subsets with greater memory potential such as stem-like memory T cell (TSCM) and central memory T cell (TCM) or terminally differentiated effector memory reexpressing CD45RA T cell (TEMRA). Additionally, the conserved lineage structure predicts that naive T cells give rise to TEM which are in a closed cycle of differentiation with TSCM and TCM. Terminally differentiated TEMRA were identified to arise from one of the three memory populations. In conclusion, we have developed a novel approach that allowed us to estimate cell kinetics and infer lineage topologies in humans in vivo and applied it to two major immune cell lineages.