Development of methods that combine biological pathway information with the analysis of genom wide association studies

Coordinator:		Prof. Dr. Roland Eils
Institution:		DKFZ Heidelberg
Homepage:		www.dkfz.de/tbi/people/

Genome wide association studies (GWAS) will be conducted within this IG for three complex genetic diseases: bipolar disorders (BPD), unipolar/major depression (UPD) and schizophrenia (SCZ). Complex diseases are a challenge for statistical analyses due to the multiple testing problem that arises when combinations of markers (gen-gen-interactions) are tested. Due to the testing of marker combinations the number of tests grows enormously. Using typical significance thresholds for each test without taking the multiple testing into account results in too optimistic results producing many false positive results that are not biologically significant.
In this subproject, we will develop methods to find multimarker combinations in genotyping data that show combined significant association with disease risk. We will search for such multimarker combinations in biologically defined functional modules. We will develop and adapt group testing methods that test for entire groups of markers instead of treating each marker independently. Groups will be defined by functional information of genes to which these markers are mapped. We will further search for combinations of markers in pathways for which independent evidence shows their involvement in pathomechanisms. Markers mapping to such pathways will be re-weighted to have higher chance to find them together in an analysis. The methods from this subproject will be applied to genotyping data from bipolar disorders, unipolar disorders/major depression and schizophrenia to detect both functional modules/pathways and multimarker combinations with strong association with disease risk, based on hypotheses on influence of functionally related groups.