The outcome of metabolomics studies strongly depends on annotations of quantified metabolites. As non-targeted procedures detect metabolites with initially unknown structures and even targeted metabolomics includes compounds with ambiguous labels, for further usage metabolites need to be characterized. In this work, I introduce two novel systems biology methods that are, unlike existing approaches, not based on fragment spectra, but primarily use the correlation structure of measured concentrations in samples of large cohorts to (i) determine main constituents of measured metabolite sums and (ii) characterize unknown metabolites including selection of candidates.
«
The outcome of metabolomics studies strongly depends on annotations of quantified metabolites. As non-targeted procedures detect metabolites with initially unknown structures and even targeted metabolomics includes compounds with ambiguous labels, for further usage metabolites need to be characterized. In this work, I introduce two novel systems biology methods that are, unlike existing approaches, not based on fragment spectra, but primarily use the correlation structure of measured concentrati...
»