PhD Defense by Anders Bilgrau

Reproducibility and Data Integration in High-Dimensional Statistics With Applications in Molecular Cancer Biology


09.11.2015 kl. 13.00 - 18.00


Reproducibility and Data Integration in High-Dimensional Statistics With Applications in Molecular Cancer Biology

High-dimensional data and statistics permeate modern science and technology. Today's biotechnology measures millions of genetic markers and provides an unprecedented detailed view of human diseases. In clinical research, such high-dimensional data has brought a yet unfulfilled promise of personalized medicine, tailor-made drugs, and clinical decisions based on our genetic makeup. The analyses of these data-intensive experiments are among some of the biggest challenges in science. From a statistical perspective, high-dimensionality is accompanied by a host of theoretical, technological, and practical problems and pitfalls, which may lead to irreproducible results and invalid conclusions. Indeed, concerns of irreproducible findings have been raised and the overlap between studies described as disappointingly small. The thesis explores statistical models to alleviate some of these issues.

Genetic studies are costly and consequently often have small sample sizes. A straightforward way and primary strategy of the thesis is to increase the reproducibility by integrating data from multiple experiments. Another contributor to the poor reproducibility is the common gene-by-gene screenings, which too simplistically consider genes to be independent. However, genes are highly dependent, operating in tightly regulated networks. Statistical network analysis is known to be highly unstable and thus often irreproducible. To this end, methods are proposed for identifying genes and gene networks whilst integrating information across multiple datasets.

Bedømmelsesudvalget består af: professor Jesper Møller (formand), Aalborg Universitet, Adjunkt Kasper Daniel Hansen, John Hopkins University, og Professor Niels Richard Hansen, Københavns Universitet.

Ordstyrer: Lektor Søren Højsgaard, Aalborg Universitet,
Vejleder: Professor Martin Bøgsted og lektor Svante Eriksen, Aalborg Universitet

Efter forsvaret vil instituttet være vært ved en lille forfriskning på instituttets adresse

Fr. Bajers vej 7 G i studenterkaffestuen.
Alle interesserede er velkomne!


Department of Mathematical Sciences


Kroghstræde 3, room 1.119

Gå til arrangementslisten


Se listen