This expansion has been made possible by major advances in the field’s two workhorse technologies: nuclear magnetic resonance, or NMR, and mass spectrometry. Both techniques can reveal information about the mass and structure of individual molecules, as well as the composition of complex molecular mixtures. Over the past couple decades, these machines have become significantly more powerful, capable of detecting more metabolites in a sample, while requiring smaller sample volumes.

But the human metabolome has remained a relative scientific frontier. Unlike in genetics, where efforts such as the Human Genome Project led to vast libraries of freely accessible data as early as 2003, scientists have had few resources to make sense of metabolites. The equipment necessary to measure and analyze them is large and expensive, and the resulting data streams can overwhelm even the best-equipped lab. Only in the past five years have scientists begun to piece together a roadmap, assembling databases of known metabolites to aid researchers in making sense of their data.

One of first researchers to join in that quest was John Markley, a biochemistry professor and director of the Nuclear Magnetic Resonance Facility at Madison (NMRFAM). Housed in the basement of the Biochemistry Addition, NMRFAM looks like a set from a James Bond movie, a vast, hangar-like room lined with gleaming domed machines. That equipment offers researchers the power and sensitivity to break a sample of blood or urine down into a roster of metabolites. Recognizing this unique capacity, Markley applied in 2004 for funding from a special National Institutes of Health Roadmap initiative called Metabolomics Technology Development to begin building tools to advance the field.

“We proposed that one of the major roadblocks in the field was the lack of a database containing data about pure, bona fide metabolites, as well as a lack of methods to rapidly collect and analyze data,” says Markley. “So that’s what we’ve been doing ever since.”

NMRFAM has now run more than 700 pure metabolites through its machines, compiling the data in a free, online database. Scientists are beginning to use the data—and NMRFAM’s technology—for a range of applications that extend well beyond human health. The aim of one of the facility’s projects is to compile a database of all the molecular constituents found in the plant cell wall, to aid researchers trying to unlock new forms of renewable energy from plants.

“Our major emphasis has been to get the technology in hand and get our database set up,” says Markley. “What excites me now is being able to apply the technology that we’ve developed to studies that are well-defined, and where we can use this approach to get solid information.”

Assadi-Porter’s PCOS project is the second such study to emerge from Markley’s lab. She first explored the power of metabolomics to monitor the progression of sepsis, a type of bacterial infection that sparks a dangerous, whole-body inflammatory response. She chose sepsis because the current testing technology is woefully inadequate. “By the time a doctor determines a person has sepsis,” she explains, “they are on the knife’s edge.”

With animal sciences professor Mark Cook and zoologist Warren Porter, Assadi-Porter began analyzing metabolites in the breath associated with sepsis. In experiments with mice, the team was able to detect sepsis two hours after the onset of infection, hours earlier than previously possible. They later found the same results in rats and chickens. The team patented its “breathalyzer” technology and then founded a medical devices company to develop it into a viable product for at-risk patients.

Assadi-Porter’s sepsis project highlights one of the main advantages of metabolomics: its acute sensitivity to what’s happening inside the body at a given moment. And that plasticity has some scientists saying that metabolomics could turn out to be the missing link in delivering on the promise of personalized medicine.

The idea at the root of personalized medicine is that every body functions a little differently, and what works for one may not at all work for another. Many people believed that the sequencing of the human genome would unlock this great vault of individuality, yielding a master guide that would tell us how to diagnose conditions and prescribe therapies that optimally fit each person’s unique genetic makeup. But while genes reveal a surfeit of information about inherited conditions—such as a patient’s predisposition to breast or colon cancer—most of our day-to-day maladies are not hard-wired into our genetic code. To get a complete understanding of the processes that govern our bodies, we need to look not just at our genome, but at the other –omes: the transcriptome, which describes all of the protein-encoding RNA molecules in our cells; the proteome, our complete set of proteins; and the metabolome.