Epigenomics

Our current research efforts are focused on machine learning-based epigenomic tumor diagnostics. The introduction of a brain tumor DNA methylation classifier (Capper et al., Nature 2018) has prompted efforts to establish a respective diagnostic pipeline in our institute. Employing DNA microarrays, we have so far analyzed DNA methylation profiles of more than 2000 tumor samples in a diagnostic context. However, as this approach is time-consuming, we have more recently begun to establish a nanopore sequencing-based pipeline to enable same-day epigenomic (and genomic) profiling. Its underlying workflow (Euskirchen et al., Acta Neuropathol 2017) is much less equipment- and resource-dependent, and allows molecular tumor typing even before histological sections or special stains become available. In addition, we recently launched an open-source portal for colleagues around the world to compare their methylation datasets with a continuously growing dataset of annotated reference cases (www.epidip.org). Finally, as part of an international consortium, we have contributed (and continue to do so) to a number of collaborative efforts; relevant publications below.