Chronic wounds are among the most disabling complications of diabetes and remain a leading cause of non-traumatic lower-limb amputation. Wound repair proceeds in a hypoxic tissue microenvironment, and the response of infiltrating immune cells to that environment is a determinant of healing outcome. This project characterises hypoxia-associated states in circulating and wound-resident immune cells in diabetes, using single-cell multi-omics. The project is carried out at the Department of Molecular Medicine and Surgery, Karolinska Institutet, Solna, in the group of Sergiu-Bogdan Catrina. Three aims structure the work. Aim 1 characterises cell-type-specific transcriptional, chromatin-accessibility and surface-protein responses to intermittent hypoxia in peripheral blood mononuclear cells (PBMCs) from patients with long-duration type 1 diabetes and matched healthy controls. Aim 2 examines immune populations in DFU tissue from patients with type 1 and type 2 diabetes, and compares their profiles with those obtained from circulating cells. Aim 3 addresses the effect of chronic hyperglycaemia on immune-cell regulatory programmes, hypoxia-responsive pathways and wound-repair-associated functions. The computational work proceeds in two phases. The first phase, for which the present allocation is primarily requested, reanalyses publicly available single-cell datasets deposited in the Gene Expression Omnibus, covering single-cell transcriptomics and spatial transcriptomics. We will assemble a harmonised reference of circulating and wound-resident immune populations across independent studies of diabetes, hypoxia exposure and chronic wounds, unify cell-type annotation across datasets, platforms and modalities, and derive reference gene and regulatory-element modules against which data generated in the second phase will be scored. All data used in this phase are de-identified and openly available; no personal data are processed under this allocation. The second phase will analyse data generated in-house from patient and control PBMCs under controlled intermittent hypoxia, comprising CITE-seq and single-cell ATAC-seq, together with single-cell data from patient of impaired diabetic wound healing. CITE-seq quantifies surface protein epitopes alongside the transcriptome in the same cells; scATAC-seq profiles chromatin accessibility, allowing transcriptional states to be linked to the underlying regulatory landscape. Human data carrying personal information will be processed on a NAISS resource approved for sensitive data under a separate allocation; only de-identified and aggregated results will be handled here. Analyses are implemented in Python (scanpy, scvi-tools, muon, snapATAC2, anndata) and R (Seurat, Signac, Bioconductor). Cross-dataset and cross-modality integration relies on variational autoencoder models (scVI, totalVI, PeakVI, MultiVI), which require GPU acceleration at the cell numbers involved. Read alignment and quantification, peak calling, quality control, doublet detection, differential accessibility and expression testing, motif enrichment and pathway analysis are CPU-bound and parallelised across samples.