GROUP: Computational Oncology

We are adapting methods from machine learning to various data types from molecular genetics. We are developing new methods for their analysis, in particular with applications to oncology.


The Computational Oncology Group focuses on the extraction of biologically or clinically relevant information from high throughput data. Currently we analyze mainly data from next-generation sequencing (NGS) and microarray experiments; the main data sources are the three German International Cancer Genome Consortium (ICGC) projects as well as the Heidelberg Center for Personalized Oncology (HIPO). To analyze these big data sets we develop and apply various statistical methods and machine learning techniques. Our goal is an improved understanding of cancer biology (this can mean the identification of driver mutations or pathways, or unraveling genotype – phenotype correlations) and findings which can be translated back into clinics, for example biomarkers for patient stratification, and actionable mutations for targeted therapies.

 

Selected projects and topics

Cancer Genome Bioinformatics  

ICGC

Our group is part of all three German projects within the ICGC. We perform variant calling from whole genome sequencing data, analysis of RNA-seq, miRNA-seq and whole genome bisulfite sequencing data and integrative analysis of the results obtained from the different data levels.

ICGC PedBrain

logo pedbrain  tumorThe PedBrain Tumor Research Project is the first German contribution to the ICGC. It focuses on medulloblastoma, pilocytic astrocytoma, and glioblastoma multiforme. In total, the PedBrain consortium aims at whole genome sequencing of more than 500 tumor / normal pairs. With the results of the study we expect to provide the basis to develop new therapies, stratify patients for risk adopted treatments, and provide markers that predict the response to the applied therapy.

ICGC MMML-Seq

MMMLseqThe ICGC-MMML-Seq (Molecular Mechanisms in Malignant Lymphoma by Sequencing) Consortium analyzes germinal-center derived B-cell malignant (non-Hodgkin) lymphomas (GCB-lymphomas). The core subtypes of GCB-lymphoma analyzed are follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma/leukemia (BL) with clinical or histologic variants or grey-zone forms to be added. Overall, 250 lymphomas and matched normal tissue will be sequenced.

ICGC Early Onset Prostate Cancer

EOPCProstate cancer is generally considered a tumor of elderly men. However, a fraction of prostate cancers are diagnosed at the age of 50 years or less.  The ICGC “Early Onset Prostate Cancer” consortium specifically analyzes prostate cancer from patients in the age group below 50 years. The aim of the project is to identify genetic and epigenetic alterations that correlate with early tumor formation.

ICGC PanCancer

The ICGC PanCancer study will reanalyze whole genome sequencing data from ~2500 tumor – normal pairs from > 25 cancer types, which was acquired in the different ICGC projects. In addition, for several tumor samples transcriptome and methylome data will be available.Together with the Korbel group at EMBL we provide one of the unified variant calling pipelines used in this project (the other pipelines come from the Sanger institute and the Broad institute).  

Personalized Oncology

We provide various levels of data analysis for several projects within the HIPO. The mid- to long-term goal of HIPO is to develop a program for personalized oncology by translating latest research and technologies from the field of functional genomics and systems biology into clinical practice. Individual cancer genome analysis will form the basis for personalized cancer treatment; pilot projects to explore these possibilities have already been started.