In kidney transplantation, deteriorated progression of rejection is considered to be a leading course of postoperative mortality. However, the conventional histologic diagnosis is limited in reading the rejection status at the molecular level, thereby triggering mismatched pathogenesis with clinical phenotypes. Here, by applying uniform manifold approximation and projection and Leiden algorithms to 2,611 publicly available microarray datasets of renal transplantation, we uncovered six rejection states with corresponding signature genes and revealed a high-risk (HR) state that was essential in promoting allograft loss. By identifying cell populations from single-cell RNA sequencing data that were associated with the six rejection states, we identified a T-cell population to be the pathogenesis-triggering cells associated with the HR rejection state. Additionally, by constructing gene regulatory networks, we identified that activated STAT4, as a core transcription factor that was regulated by PTPN6 in T cells, was closely linked to poor allograft function and prognosis. Taken together, our study provides a novel strategy to help with the precise diagnosis of kidney allograft rejection progression, which is powerful in investigating the underlying molecular pathogenesis, and therefore, for further clinical intervention.