Multiplexed imaging analysis of human pancreatic islets from donors with and without type 2 diabetes

This dataset is part of the manuscript Genetic risk converges on regulatory networks mediating early type 2 diabetes (Walker, Saunders, & Rai et al., 2023), a body of work that includes tissue imaging, sorted islet cell transcriptomics, and islet functional analysis of donors with early-stage type 2 diabetes (T2D) and control donors. Raw imaging data was acquired from 4% PFA-fixed, cryopreserved human pancreas tissue using the CODEX system (now PhenoCycler Open; Akoya Biosciences) integrated with a BZ-X810 epifluorescence microscope (Keyence) with a CFI plan Apo I 20x/0.75 objective (Nikon). Image alignment, stitching, background subtraction, and deconvolution were performed using the CODEX Processor v1.7.0.6 (Akoya Biosciences). Tissue and islet areas were annotated by hand to exclude out-of-focus regions and poor tissue quality. Islets (estimated diameter ≥50 μm; mean 42 islets/donor) were annotated based on DAPI and CHGA channels. Cell segmentation and cell type annotations were performed using the HALO HighPlex FL v3.2.1 module (Indica Labs). For cell neighborhood (CN) analysis, two methods were applied in parallel to CODEX data from annotated islets: a community detection method, termed Dynamic CF-IDF, and a k-means approach. Processed imaging data is available via Pancreatlas (RRID:SCR_018567); packages used for cell neighborhood analyses are published in Github.