Signal Strength Aware Latent Spaces Reveal Molecularly Distinct Substructures within Human Kidney Tissue

This data repository contains data required to run one of the example notebooks in the sisal package. The SiSAL package is an approach based on the beta-variational autoencoder and kernel density estimation to dissect data along independent, uncertainty-aware, and interpretable (yet non-linear) latent axes. It includes a novel comparative-latent-traversal algorithm to translate latent findings back into the original measurement context. You can find demonstration in the /experiments folder. It includes an imaging mass spectrometry-based molecular imaging of human kidney and a synthetic dataset. The approach’s disentangling properties are shown to impress a latent space structure that separates signal strength from relative signal content, offering exceptional chemical insight. Data included: centroids.h5 - A HDF5 file containing ion images in a matrix format. The ion images correspond to the set of images presented in our HuBMAP Kidney Atlas. normalizations.h5 - A HDF5 file containing the normalization factors that can be applied to the ion images. image_info.npz - a compressed numpy (Python) file containing information about the ion images. Metadata such as image shape or the total number of pixels in a dataset Collecting_duct.h5 - A HDF5 file contianing mask of the Collecting Duct. Descending_Thin_Limb.h5 - A HDF5 file contianing mask of the Thin Descending Limb. Distal_Tubule.h5 - A HDF5 file contianing mask of the Distal Tubule. Glomerulus.h5 - A HDF5 file contianing mask of the Glomerulus. Proximal_Tubule.h5 - A HDF5 file contianing mask of the Proximal Tubule. Thick_Ascending_Limb.h5 - A HDF5 file contianing mask of the Thick Ascending Limb. For information about the data collection, please refer to the HuBMAP Kidney Atlas.