Can we find disease fingerprints in our metabolome?

Background: The regulation of all processes taking place in a human organism is known as metabolism. Presence and concentration of metabolites, small molecules that take part in metabolism, is influenced by genetics and the environment. Many complex diseases also have a genetic and environmental background and influence on the concentration of metabolites, which may, therefore, be used as biomarkers for disease detection. Acquiring nuclear magnetic resonance (NMR) metabolomics data is relatively easy and has a low cost compared to other techniques used to characterize the metabolome. Yet, analysis of such NMR-based metabolomics data requires a sophisticated computational pipeline, for extracting information relevant to disease, like biomarkers.

Goal: The aim of this project is to write R code to analyse NMR data from complex biological samples. We aim to extract information from urinary NMR data regarding a disease of interest and if possible identify metabolites as biomarkers. We will use data obtained from CoLaus, a longitudinal study investigating the prevalence of cardiovascular disease risk factor in the population of Lausanne.

Tools: Using biclustering methods students will be asked to analyze urinary NMR metabolomics data, aiming to identify a group of individuals with a potentially common disease background. Moreover, students will use linear regression analysis to identify metabolites related to a disease of interest.