In this project, we will leverage data from several in-house cellular drug response screens involving hundreds of approved drugs and hundreds of cell lines as well as proteome expression data collected for the same systems to test various ML and AI strategies for drug response prediction. The resulting prediction models will be tested in silico using publicly available data, experimentally verified via focused laboratory experiments in cell lines, organoids and mouse models and applied to data collected for cancer patients in the NCT/DKTK-MASTER (Molecularly Aided Stratification for Tumor Eradication) trial. As output, this project will (i) offer unified access to comprehensive human and mouse drug response screens including genomics, transcriptomic and proteomic profiles, (ii) develop a series of unique and complementary state-of-the art AI / ML methods for drug response prediction, (iii) verify the resulting drug response prediction models via application to independent data and wet lab experiments as well as to investigate their clinical potential. By covering the whole chain from data curation, method development towards experimental and clinical application as well as making all data and tools available as part of the existing ProteomicsDB platform, we will help in advancing systems medicine in general, and the work of molecular tumor boards in particular.