Metabolite profiles accessible through modern metabolomics approaches provide readouts of a person's physiological state at a very high "metabolic resolution", covering a broad range of hundreds of metabolites (small molecules) in biofluids such as blood or urine. Recent advances enabled this deep metabolic phenotyping on an epidemiological scale, opening new avenues to study human metabolism and its variation in health and disease based on large population studies.

Extrinsic factors such as fasting status, physical activity and circadian rhythm are known to strongly influence the levels of numerous metabolites in human biofluids (Krug et al., FASEB, 2012; Raffler et al., in preparation). Despite these dynamic changes, various studies demonstrated that a person’s metabolome, which is also influenced by genetic variation (Shin et al., Nature Genet., 2014), is an individual characteristic and usually is very stable over time (Yousri et al., Metabolomics, 2014). Changes in the personal metabolome have even been linked to worse health outcomes (Lacruz et al., Scientific Rep., 2018), making metabolomics a promising tool for capturing and monitoring individual metabolic health.

In this seminar, I will show examples for the variability and the stability of metabolomes taken from large-scale longitudinal metabolomics studies. I will introduce various biostatistic and bioinformatic approaches used to analyze these data, to integrate them with other omics layers, and to support their biological interpretation, going from visual data exploration over association analyses to network-based techniques.