DeepDoublet identifies neighboring cell dependent gene expression

ADoub_new.zip contains ADoub.npy and ADoub_meta.csvADoub.csv has the UMI counts of 8000 genes of 400,000 artificial doublets.ADoub_meta.csv has the meta information of the 400,000 artificial doublets.The meta file has four columns:| Column | Description ||---------|--------------------------------------------------|| `s1` | Index of the first single cell || `MF_s1` | Proportion of the first single cell (`s1`) || `s2` | Index of the second single cell || `MF_s2` | Proportion of the second single cell (`s2`) |How to load the data?ADoub_meta = pd.read_csv('ADoub_meta.csv')X = np.load('ADoub.csv')Trained modelsDecomp_hep_LEC_1.h5 is the model to decompose the doublets to hepatocytes.Decomp_hep_LEC_2.h5 is the model to decompose the doublets to LECs (liver endothelial cells).