SCENE Joint Embedding of Cells and Genes

Fit SCENE (Single-Cell Euclidean Network Embedding, Møberg et al. 2026, MIT) on a single-cell RNA-seq dataset: cells and genes land in one Euclidean space where the cell–gene distance enters the count model directly, so proximity means expression and the space can be read off a plot. Runs on CPU in the Cortexa sandbox on an .h5ad with raw UMI counts (public URL or Library file), subsampled to about 10k cells and 2k highly variable genes. Returns nearest genes per cell population, nearest populations per gene of interest, cell and gene coordinates as CSV, a joint figure and a replayable code cell.