there should be a mechanism for reading from the weight_spectrum column in more recent datasets
there should be a mechanism for reading from the weight_spectrum column in more recent datasets