computational
modeling in biology

Font size »A . A+ . A++ .

Metabolomics

The CMB group is involved in several metabolomics projects at the Helmholtz center, covering the systems biology and modeling aspects of this newly arising -omics technology.

Main contacts at CMB: Jan Krumsiek, Jörg Bartel, Ferdinand Stückler

Gaussian graphical models

  • We established Gaussian graphical models, which are based on so-called partial correlation coefficients, as a useful tool for the analysis of high-throughput metabolomics data.
  • Generally, reconstructed edges in the network correspond to metabolite pairs that share direct connections in the metabolic pathway.
  • Collaboration partners: Gabi Kastenmüller, Karsten Suhre, Thomas Illig, Jerzi Adamski. Our study is based on data from the KORA population cohort. 

 

Network-based analysis of metabolomics data

  • The GGM method has been used e.g. to elucidate sex-specific metabolite differences in the KORA population cohort.
  • We combine the results from differential statistical analysis with metabolic networks in order to find regulated pathway regions. 
  • Collaboration partners: Gabi Kastenmüller, Thomas Illig, Karl-Heinz Ladwig, Jakob Linseisen

 

Integration of metabolomics with transcriptomics/proteomics

  • We integrate metabolomics data with transcriptomics and proteomics data, both on a data-driven and on a knowledge-driven basis. 
  • For the data-driven approach, we currently estimate Gaussian graphical models on joint data matrices containing multiple omics datasets.
  • On a knowledge-driven level, we project the differential activity of measured entities onto metabolic pathways and apply specifically designed algorithms to find enriched regions (see image).
  • Collaboration partners: Martin von Bergen, Holger Prokisch, Susanne Neschen / Martin Hrabé de Angelis

 

Modeling mitochondrial beta-oxidation

  • We created a simple mathematical model of mitochondrial beta-oxidation, that is the catabolic breakdown of fatty acids.
  • Fitting the model to measured data points yields relative reaction rates for each breakdown step in the beta-oxidation cascade.

  • Relative reaction rates could be applied to better explain interindividual variation in metabolic patterns and improved the characterization of distinct metabolic phenotypes.

  • This work is part of the "Munich Functional Metabolomics Initiative"  MuFuMet



References