Data corpus for a gene representation framework of spatial transcriptomics

The data corpus was used to implement a gene representation framework for spatial transcriptomics. The corpus consists of two Visium datasets after cell-type deconvolution with Cell2location. One dataset comprises 8 colon and terminal ileum tissue sections (.h5ad) from human adult and fetal (ref 1). The other dataset comprises 28 heart tissue sections (.h5ad) from human normal myocardium and multiple physiological zones at distinct timepoints after myocardial infarction (ref 2). A total of about 580,000 spatially resolved cells are in the corpus.