We have earlier been focusing on measurement error problems in traditional epidemiological situations. Statistical genomics is a rather new area of research where much efforts and resources are directed. Measurements of gene expressions and products of gene expressions, like protein levels, are noisy.
Lately, we have been focusing more on problems related to measurement error and measurement uncertainty in this area, and we are at the moment working on projects related to measurement error in regression methods dealing with the p>>n problem (lasso, pcr).
Another problem in genomics that has received some attention during the last few years, is how to integrate data from different sources / different layers of genomic information (SNP’s, copy number variation, expressions …) into one coherent, analytic framework. One approach here is through variants of Principal Component Analysis (PCA).