RIDDEN Model: Receptor and Ligand Perturbation Signatures for Model Development

The RIDDEN (Data-driven Inference of Receptor Activity from Transcriptomic Data) gene expression dataset was created using data from the LINCS L1000 database and curated ligand-receptor interactions from the OmniPath resource. This dataset includes 14,463 perturbation profiles for 229 receptors across 228 cell lines. The perturbation signatures consist of genetic perturbations (CRISPR and shRNA and overexpression) and chemical perturbations (ligand stimulation and drug treatment) of receptors and their interacting ligands. Curated ligand-receptor interactions were queried from OmniPath. The resulting dataset includes receptor and ligand perturbation signatures and a table of the prior knowledge of curated ligand-receptor interactions that serve as a basis for the receptor-activity predicting RIDDEN model.