Software and Science

A blog on Metabolomics, Algorithms, and Software Development.

Metabolomics is a rather young life sciences discipline, although its origins go back to more than a century of research in biology and biochemistry.

Studying the Metabolome

Metabolomics is all about metabolites, which are the entirety of small molecules that are digested, secreted, excreted, inhaled and exhaled. They build the foundation of every living organism on earth. They deliver the parts necessary to build nucleotides that make up your DNA, peptides, that make up your proteins, fat that keeps you warm, sugars that make you run, and much much more.

In order to study an organism’s metabolome, sophisticated methods are required, both in the experimental setup, trying to capture the organism’s adaptation or reaction to imposed stimuli, and in the downstream analysis of the data gathered from the experiment. Complex organisms usually do not lend themselves to an immediate examination of their complete metabolome. In fact, the metabolism differs largely, depending on the specific locus within a (complex) organism that consists of different tissues and various cell types.

Separation and Detection

Owing to the large diversity of metabolites, no single analytical method is able to measure and quantify all of them in one go. In practice, multiple analytical methods need to be combined to cover the potential size range of metabolites, starting with smaller, volatile compounds that can be detected by gas chromatography-mass spectrometry. Larger compounds may be detected by liquid chromatography-mass spectrometry or other separation methods, like ion mobility coupled to mass spectrometry.

Data Processing

The vast amount of data that is generated from experiments that try to determine metabolite abundances and identities in hundreds or even thousands of samples requires sophisticated processing methods with a solid foundation in computer science and statistics. These methods cover diverse areas from signal processing and filtering, clustering and machine learning, and statistical comparison and significance testing.


Due to the diverse collection of methods, many different disciplines are involved when working on metabolomics experiments. Metabolomics is thus at its heart an interdisciplinary area of research and requires a common vocabulary for efficient communication between researchers with different backgrounds.