Doctor Oliver Fiehn

"Our laboratory is committed to understanding metabolism on a comprehensive scale," Fiehn says. "We develop informatics tools, and we apply novel types of separation and mass spectrometry to get a very broad view of what is going on in biological systems."

The end game is twofold, he says. One is understanding the mechanisms of nature; the other is translating that knowledge into early diagnostic tests for various human diseases.

"Mass spectrometers like the ones from Agilent—the GC/MS as well as the LC/MS systems—get ever-better in terms of sensitivity and the ability to see signals that can be applied to biomedical research, like predicting who will get a heart attack in the next two years or two months," he says.

Fiehn believes that the research he and others are doing will someday lead to early and very specific diagnoses.

Fiehn and others are laying the groundwork by gathering as much mass spectral data as possible and putting it into databases that the scientific community can access—either freely in the case of government-funded research, or for a fee in the case of company-funded work like the Fiehn-Agilent GC/MS libraries.

"Last year, we published LipidBlast, a database of 200,000 mass spectra of complex and neutral lipids, representing 29 lipid classes," he says.

Fiehn acknowledges, however, that it's just a start.

"While there is primary metabolism—that is, sugars, amino acids, and hydroxyl acids that are commonly found in many species—the real diversity in small molecules comes from the plethora of terpenoids, lipids, flavonoids, or phenol-derived compounds, which are so abundant and so manifold that no single laboratory and no single library can have all of them," Fiehn says.

"So we need to predict the spectra. If we really want to understand nature, we have to be able to predict mass spectra. That is something that is a years-long undertaking that we have started together with the community at large. We are grateful when Agilent and others support these efforts."

To make such predictions, scientists draw on existing data and first principles of physics.

"Once you know how specific molecules fragment, you can make chemical scaffolds in the computer and say, ‘OK, so let's imagine instead of this acyl chain we have a longer acyl chain, or instead of this molecule we have the methylated variant of the molecule. How would that affect the mass spectra in terms of the ions you see and in terms of abundances?'" he says. "The more mass spectra of known compounds we have, the better this model building."