Breast cancer IMC dataset for InterSTELLAR training

graph_data.npy: preprocessed Breast cancer IMC dataset for InterSTELLAR training, saved as a .npy file.

1. Read this file with python and get a list named all_graph.  This list includes 368 elements. The first 366 elements are the information from each tissue, the last two elements are the mean and standard deviation of the cell feature data after log-transformation with shape 30x1.
2. Each sub-element all_graph[i] is alone a new list with three elements, including cell feature matrix all_graph[i][0], cell locations all_graph[i][1], tissue labels all_graph[i][2] and cell phenotypes all_graph[i][3].
3. all_graph[i][0] is a Nx30 matrix, corresponding to 30 cell markers;  all_graph[i][1] is a Nx2 matrix, corresponding to x and y locations of a single cell;  all_graph[i][2] is a list with 5 elements, corresponding to tissue phenotypes (0: healthy, 1: TNBC, 2: Non-TNBC cancers), tissue area (um²), tumor grade, overall survival time (month) and  patient status (alive or death);  all_graph[i][3] is a list with N elements corresponding to the cell phenotypes.

cell_masks.npy: the cell segmentation maskes corresponding to the 366 tissues.