ABSTRACT
The goal of data fusion in metabolomics is to combine data of various platforms measured on the same set of samples or individuals. In metabolomics research, multiple metabolomics platforms are often used to screen for differences between the samples. Also metabolomics measurements can be combined with transcriptomics and proteomics measurements of the same samples or the same individuals. Other applications combine metabolomics measurements of multiple compartments within the same individual. Data fusion methods can be used to explore relationships between features of different sets of data and to find what the data sets have in common and what is distinct in the separate data sets. In this chapter we review correlation based methods, simultaneous component analysis based methods and methods that distinguish between common and distinct variation.
