Beyond TCR clonotypes: integrating transcriptomic similarity and V(D)J features to infer inter-tissue naïve T-cell dynamics

The thymus is the site where bone marrow-derived precursors differentiate into mature T cells before entering peripheral blood, a process called thymopoiesis. Although thymopoiesis declines with age, it may be reactivated in adults with certain thymic epithelial tumors (TETs), particularly type AB and B thymomas. Here, the neoplastic thymus may harbor developing T cells alongside infiltrating mature T cells, resulting in a complex pattern of T-cell trafficking. We performed scRNA-seq paired with αβ and γδ TCR-seq on matched thymic tissue and blood from 19 TET patients. Integration with public healthy datasets yielded 526,579 cells. Using Scirpy, we identified TCR clonotypes shared between thymic and circulating T cells: however, this analysis cannot capture relationships among non-clonally-expanded naïve T cells. To overcome this limitation, we leveraged Dandelion VDJ feature space to extend TCR similarity beyond canonical T-cell clonotypes. In parallel, we developed a cross-compartment transcriptional similarity score based on neighborhood overlap in the scVI latent space. Together, these complementary approaches enabled us to resolve thymic–circulating relationships between newly generated and pre-existing T-cell populations beyond canonical clonotype sharing, providing a new framework to investigate inter-compartment naïve T-cell dynamics.