"It is well documented that each human cancer has acquired a large number of mutations (50-500), some of which are drivers for tumorigenicity, others of which are indolent passengers. It has also been established that such mutations can be recognized by the patient´s T cells via the presentation of mutated peptides on HLA molecules of a tumor cell. If such immune responses against cancer mutations are generally relevant for cancer development (according to Paul Ehrlich´s hypothesis), one can predict several consequences: 1) The spectrum of mutations found in an established tumor of a patient should be diminished for those that give rise to peptides presented by the HLA molecules of this patient, but not for peptides relevant for other HLA molecules. 2) In HLA-negative tumor cells, the spectrum of mutations (those that occurred after HLA loss) should no longer be dependent on the patient´s HLA molecules. 3) A cancer patient should have T cells against mutations that occurred at an early point in cancer development but that are no longer present in the tumor at the time of diagnosis because of negative selection of such tumor cells by the immune response. The aim of this proposal is to test the hypothesis by studying these three predictions in a large cohort of cancer patients treated at the University Hospital of Tübingen. Clinically relevant human cancer samples from several cancer entities will be tested for mutations by exome and transcriptome sequencing. Those mutations actually presented as HLA ligands will be analyzed by mass spectrometry, and the patient´s T-cell repertoire against mutations will be analyzed. The result of this research should be the confirmation, the rejection, or an amendment of the hypothesis. The outcome will be of decisive influence on the cancer therapy of the future, both for the targeting of cancer pathways using small molecule inhibitors, and for the individualized immunotherapy of cancer by targeting mutations."