T cells develop in the thymus, where they acquire a range of distinct functional identities. Studies over the years have defined, at a population level, many of the specific genes that have to be activated or silenced as T cell progenitors differentiate toward either the αβ (alphabeta) or γδ (gammadelta) lineages. However, it is still unknown whether this combination of activation/silencing actually occurs in each and every cell as they differentiate. To investigate this, single cell RNA sequencing (scRNAseq) technology has been employed to re-assemble de novo a model of the early stage in T cell development based on the transcriptional profiles of individual cells. Two sets of cells were analyzed using Chromium 10x Genomics scRNAseq. A pure pool of CD4-CD8- double negative (DN) and γδ thymocytes were sorted for the first run, which yielded 5,527 high quality transcriptomes. For the second run, only DN1 and DN2 thymocytes were analysed, which generated 8,837 high quality transcriptomes. Hierarchical clustering identified 19 and 26 sub-populations within the first and second run respectively. This included 8 different populations that fall within the previous DN1 definition indicating much more complex than standard view of early T cell development. Assembly of these novel thymocyte populations into developmental trajectories based on gene expression revealed key findings. It indicated that the decision to differentiate into αβ or γδ T cells occurs at a much earlier stage than the current model and that there are distinct DN2 sub-populations and potentially DN1 that are specific to only one of the two developmental pathways. Analysis in the OP9-DL1 cultures confirmed that specific sub-populations of DN1s and DN2s differentiated into only αβ or γδ thymocytes. This is consistent with the αβ or γδ lineage decision being determined from the earliest stages of T cell development.