Like in the example code, if I first perform a filter, like selecting the protein_coding gene, and then normalize it using raw_normalized_by_feature_length(), then the shape will not be broadcastable.
dX = dX.filter(explorer.make_filter(feature_types=["protein_coding"])
# common attributes are available such as tissues, assays, organism, disease, etc
tissues = dX.tissues
# .raw_normalized_by_feature_length gives you a numpy array
# .raw is also available (raw counts)
cells_by_types = [
dX.filter(
explorer.make_filter(
max_cells=5_00,
tissues=[tissue]
)
).raw_normalized_by_feature_length
for tissue in tissues
]
Like in the example code, if I first perform a filter, like selecting the protein_coding gene, and then normalize it using raw_normalized_by_feature_length(), then the shape will not be broadcastable.