Data for Scalable parametric encoding of multiple modalities

Two folders with figures created using Generative Encoding, using data from Stenbeck et al [1]:

Folder 1 - epithelial_immune_transition

This contains in-silico perturbations of epithelial and immune genes. Using generative encoders, the transformation begins with a histology to gene expression transformation, a perturbation of the gene expression vector, followed by an inverse transformation back to histology.

Folder 2 - heatmap_he_et_al

This contains in-silico transformation of histology tissue to gene expression. It is similar to the first folder, yet is only a single transformation (without an inverse) from histology to gene expression.

[1] Stenbeck, Linnea; Bergenstråhle, Ludvig; Lundeberg, Joakim; Borg, Åke (2021), “Human breast cancer in situ capturing transcriptomics”, Mendeley Data, V5, doi: 10.17632/29ntw7sh4r.5