Using Amira to manually segment organelles in vEM for machine learning v2

In this protocol we describe the voxel-based classification of organelles and cellular substructures in volume electron microscopy data used to train deep learning networks for automated segmentation. This protocol is centered around using Amira software for ‘painting’ and details how the CellMap Project Team has defined [up to] 37 cellular substructures for automated semantic segmentation tasks. This protocol was used to train the networks described in Heinrich et al, Nature (2021).