EVENFLOW Personalised Medicine use case: Clear Cell Renal Cell Carcinoma (ccRCC, aka KIRC) datasets collection

These datasets contain the synthetic data generated with a VAE from the TCGA-KIRC dataset. The file Static_KIRC.csv contains a pre-processed version of the bulk-RNASeq dataset from TCGA-KIRC. nodes_metadata.csv contains the clinical information of the patients (with columns: sample name, gender, race, and stage). A synthetic dataset is available (compressed), where 50 timepoints were inferred: trajectories_forward_test.csv.tar.gz simulates the progression of RNASeq data between patients at stages early and late. Both the static and trajectory datasets where analyzed for their biological interpretation with a Differential Expression (DESeq) analysis followed by GSEA on all the pathways available in the Reactome database. The results of the static analysis are included in static_gsea_reports_kirc.csv and the results on the trajectories may be found in trajectories_gsea_reports_kirc.csv.tar.gz.

Trained models used for the synthetic data generation are available in Hugging Face.