Computational Epigenomics

Theme picture

Computational Epigenomics is a research group within the department of Epigenomics and Cancer Risk Factors in the German Cancer Research Center, Heidelberg, Germany.

Overview

Recent advances in microarray and sequencing technologies made the genome-wide profiling of DNA methylation and histone modifications feasible and scalable to cohorts containing hundreds or even thousands of samples. Large international consortia, such as the Encylopedia of DNA Elements, The Cancer Genome Atlas Project, the International Cancer Genome Consortium and the International Human Epigenome Consortium have systematically generated comprehensive epigenome-wide datasets on tissue samples and cell lines in human. These vast cohorts open unprecedented opportunities to study the genetic regulation, intracellular heterogeneity and disease progression at the molecular level, but they also bring up enormous challenges in data processing and analysis.

Major Research Directions

Our major goal is to improve the knowledge about the role of epigenetic processes, with a focus on the initiation and progression of cardiovascular diseases and cancer. We map the complex interplay between genome, epigenome and transcriptome by designing and using novel, integrative analysis techniques and advanced visualization approaches.

gender prediction based on Infinium 450k data
sample arrangement on 96-well plate

We also design custom strategies for biomarker prioritization and validation for various purposes, including therapy suggestions in early onset prostate cancer, classification of rare tumors and non-invasive diagnosis of cardiovascular diseases. We also study the predictive potential of DNA methylation and the applicability of methylation-based prediction models in clinical settings.

To improve the reproducibility and comparability of the currently used genome-wide technologies in detecting local and global changes in DNA methylation and histone marks, we investigate the strengths and limitations of these methods and their processing pipelines.