|search externally:||Google Scholar, Springer, CiteSeer, Microsoft Academic Search, Scirus , DBlife|
The study of biological processes in the cell has been revolutionized by the advent of high-throughput techniques. Such techniques generate complex experimental data from various aspects of the cell. Sound statistical methodology for the analysis of these data is imperative for the understanding of the molecular mechanisms underlying the data. In this field I work on the design and analysis of copy number, gene expression, and increasingly on methylation and microRNA microarray experiments. In recent years my research yielded several contributions to the preprocessing and downstream analysis of DNA copy number data; a two-sample test for differential gene expression when one group exhibits heterogeneous expression levels; the comparison of survival prediction methods built from gene expression data.
A special focus of my research over the last two years has been the integrative analysis of DNA copy number and gene expression data. Among others this lead to the development of a nonparametric test for the detection of DNA copy number aberrations induced differential gene expression; the formulation of a multivariate model for regional co-expression associated with DNA copy number aberrations; and the methodology to explore the effect of DNA copy number aberrations on gene expression levels within a pathway. Stimulated by developments in the VUmc's cancer clinic, the special focus is broadening to include other genomics data like methylation and microRNA data.
Originating in my PhD-thesis I also develop statistical methodology for the evaluation of categorical (nominal, binary, ordinal) measurement systems in the absence of a gold standard. This encompasses design and sampling strategy of the measurement system analysis experiment; the study of mixture models / latent variable models (their formulation, identifiability, estimation, and diagnostics) that describe the outcome of the experiment; and the development of tangible summary statistics.
A random coefficients model for regional co-expression associated with DNA copy number aberrations
as author at Cancer Bioinformatics Workshop, Cambridge 2010,