ABSTRACT A time-sensitive window of opportunity exists in preventing and treating type 2 diabetes (T2D) in youth where the identification of new molecular targets could be used for disease prevention and control. The key objectives of this proposal are twofold. First, this proposal provides a combined strategy of didactic training and mentored training for (a) the measurement of single-cell chromatin accessibility and single-cell transcription factors associated with T2D using innovative wet-lab techniques; (b) bioinformatics, including machine learning, for the integration of multi-omic, clinical, and sociodemographic data to classify phenotypes of T2D; and (c) leadership and management skills required of high-quality clinical studies. The co-mentors will provide integrated training in single-cell epigenomic and transcriptomic lab techniques, data analyses, and coordinated training with four advisory committee experts. The second key objective of this proposal is the implementation of hypothesis- driven research aims around subtypes of immune cells that have pro-inflammatory vs. anti-inflammatory function. Single-cell sequencing will overcome the limitations of bulk epigenetic and RNA studies and improve attribution of gene regulation in individual cell types. We leverage the recruitment platform from our KL2 parent study to recruit youth and examine single-cell chromatin accessibility associated with T2D. In this cross sectional study, we will collect new peripheral blood samples from two groups of youth aged 10-20-years: new- onset T2D case participants (n=96,000 cells from 24 participants) and healthy, normoglycemia control participants (n=96,000 cells from 24 participants), obtained during a single study visit by oral glucose tolerance test. Our specific aims include (a) testing the hypothesis that T helper (Th)-17 cells will be more abundant in T2D compared to healthy control participants and regulatory T cells (Treg) more abundant in control participants, (b) testing the hypothesis that genes determining T Helper cell type, Rorc and Foxp3, will have different levels of chromatin accessibility and different gene expression in clinical groups, and (c) machine learning based on high-throughput single-cell epigenomic, transcriptomic, sociodemographic, and clinical data to classify phenotypes of T2D. In summary, this proposal assembles a team of mentors and advisors across scientific disciplines to provide thorough training in research related to single-cell epigenetics, gene expression, bioinformatics, machine learning, and leadership in clinical studies, with emphasis on the development of future research for NIH grant support. The innovations of the proposed analyses, namely the integration of single cell epigenetic and single cell gene expression, will enable construction of gene regulatory networks that may inform new strategies to protect youth from T2D and related complications.