Molecular phenotyping, in particular metabonomics and proteomics, aim to identify biomarkers that are able to predict which individuals are likely to develop a chronic disease, long before symptoms appear. This enterprise may benefit much from the development of appropriate methods of statistical analysis. A statistical challenge in both metabonomics and proteomics is to make predictions on the basis of individual-level information that is typically high dimensional and takes the form of a continuous profile (e.g. spectrum). In addition, real metabonomics/proteomics applications often involve integrating data from different sources or experimental platforms. This Course will review traditional as well as state-of-the-art statistical solutions to the above challenges. The first part of the Course will focus on the analysis of Metabonomics data and the second part on the analysis of Proteomics data. For each of these topics, the Course will provide insight into the nature of the data, an overview of experimental design issues, a discussion of data pre-processing and normalization, and an overview of traditional as well as more advanced approaches to statistical analysis. Relevant inferential issues, such as dealing with multiple comparisons, will also be discussed. The Course will include computer practical sessions during the afternoons to allow participants to apply the more traditional statistical methods in the analysis of real metabonomics and proteomics data.
The last day of the Course will be devoted to advanced, state-of-the-art, statistical methods for the analysis of metabonomics and proteomics data. A forum for discussion will be set up with the aim of promoting an exchange of ideas between the speakers and audience.
The Course is intended for an audience with some previous familiarity with statistics and data analysis. Prior knowledge of metabonomics and proteomics is advantageous but is not a pre-requisite for attending the Course.