All preprocessing details can be found in https://github.com/daisybio/preprocess_drp_dataor at https://github.com/daisybio/drevalpy Change in this version: fixed drug relations for gcmf model (identifiers are now pubchemid not names) Changes before: Added pathway activity features (pathway_features.csv) and SMILES-vector drug embeddings (drug_smilesvec.csv) to all datasets (CCLE, CTRPv1, CTRPv2, GDSC1, GDSC2, BeatAML2, PDX_Bruna). The pEC50s were shifted by -6 because of the code update -> now corrected. Meta now contains the gene lists, so they don't have to be duplicated 2 new datasets: BeatAML2 and PDX_Bruna Corrected wrong original AUC values of CTRPv2 Added MolGNet model to meta Updated gene lists: intersection gene lists now also intersect BeatAML2 and PDX_Bruna. Reduced gene lists only contain genes contained in all datasets. Removed TOYv1 and TOYv2 because they are now available from nf-core/test-datasets, and we don't want to maintain them at two sites. added drug relations for gcmf model