Diabetes mellitus (DM) is a global health challenge that affects nearly 463 million people worldwide. Type 1 diabetes (T1D) accounts for 5-10% of all diabetes cases and is increasing at a rate of 2-6% annually. Recent research has challenged the dogma that all β cells are eventually destroyed in type 1 diabetes (T1D). Emerging data suggest that some cells in long- duration disease may be protected from autoimmunity owing to the acquisition of a ‘de- differentiated' phenotype that makes them less visible to the immune system. However, a consensus definition and the precise phenotype of a de-differentiated cell has yet to be established. Moreover, the heterogeneity of such a phenotype between single cells within and between islets and persons with T1D is not clear. Using single molecule Fluorescence In-situ Hybridization, our preliminary data has revealed significant heterogeneity of the spatial transcriptome in cells. This project will utilize the newly emerging tools such as computer vision and artificial intelligence, to deepen the exploration of the FISH images of human pancreatic tissues. We will develop computational tools and bioinformatics strategies to understand the spatial distribution of the transcriptome in T1D β cell phenotypes. (1) Develop computational tools to process, analyze, and quantify multiplexed transcriptomic images and proteomic images of pancreatic tissue from human organ donors. (2) Identify representative features of RNA expression in β cell phenotypes in long-duration T1D and during T1D evolution with bioinformatics strategies. Clarifying the molecular phenotype of persistent cells in long-duration disease could have important implications for T1D therapeutics, and it has the potential to inform the development of disease modifying interventions aimed at improving the function of these cells.