Statistical methodologies for genomic research

To convert genomic data into bio-medical knowledge poses major challenges to statistical sciences. With the general aim of contributing widely applicable methodology for the study of complex genomic data, this project will help identify genotypes and gene signatures related to disease risk, disease pathogenesis and therapy efficacy for several important diseases (including breast, ovarian and cervix cancer, heart failure and preeclampsia), object of research of our partners and originating from our statis tical genomics consulting service. Characteristic features of genomic data, originating from high throughput technologies, are their extraordinary dimension, the many sources of systematic and random errors, the complex dependencies between measurements and the small number of biological replicates. This calls for new tools to analyse and merge high dimensional genomic data of different types; new methods to merge genomic data and clinical and environmental information; new methods for association studi es and SNP haplotypes analysis in genome-wide scans; better sampling designs; new approaches to pathways discovery; new techniques to analyse complex event histories that include genomic data; – all equipped with measures of uncertainty. Our research te am includes internationally established statisticians, who have contributed important ideas to modern bio- and methodological statistics at large, affiliated with the Section of Medical Statistics of the University of Oslo (UiO), the Department of Mathema tics (UiO), the Department of Informatics (UiO) and the Norwegian Computing Centre (NR). Our bio-medical partners are research groups of international reputation at the Norwegian Cancer Hospital (DNR), Ullevål University Hospital, several departments of t he University of Oslo and The National Institute of Public Health.