We study metabolism using mass spectrometry, isotope tracers, and computational modeling.  Our goal is to elucidate the full metabolic network, its activity and how this activity is regulated: a comprehensive view of metabolic function. To this end, we develop new measurement and modeling approaches. We then apply them to investigate yeast and mammalian physiology, biofuel production, diabetes, aging, and cancer.

What is the scope of metabolism?
Mass spectrometry detects more than 10,000 small molecule peaks in typical biological specimens. Of these, only a small minority are molecular ions of known metabolites. Many arise from some form of analytical artifact, but others capture genuine metabolic dark matter. We estimate that yeasts make several hundred new metabolites and mammals a few thousand. Novel mammalian metabolites likely include major players in health and disease.

We have steadily contributed to the illumination of these metabolites. In 2009, together with colleagues, we revealed that a previously obscure compound, 2-hydroxyglutarate, is a major cause of cancer. Inhibitors of the production of this oncometabolite are now FDA-approved for leukemia, bile duct cancer, and brain cancer, with multiyear disease control in brain cancer, transforming the lives of patients with this terrible disease. 

Recently, we are working closely with the chemical AI-expert Michael Skinnider to accelerate novel metabolite discovery. This integration of AI and experiments is revealing new branches of mammalian metabolism. These include reactions that are strongly perturbed in aging and likely contribute to aging-associated decline. We are actively seeking the enzymes producing these new metabolites, the cause of the age-related changes, and opportunities for antiaging and anticancer interventions.

Can we improve metabolic measurements?
Metabolism is dynamic, constantly converting incoming nutrients into energy and biomass. We helped pioneer mass spectrometry-based metabolomics, enabling simultaneous concentration measurements for 100s of known metabolites. In parallel, we innovated methods to understand metabolite production and consumption rates (metabolic fluxes).  

Unlike metabolites themselves, fluxes cannot be captured in a vial and measured in a mass spectrometer. To quantitate fluxes, we developed methods involving feeding isotope tracers and measuring the kinetics and extent of intracellular metabolite labeling. Labeling patterns are then computationally integrated to yield fluxes. We have expanded our toolbox for flux quantitation from cultured cells to mammals. Currently, we continue to develop new flux measurement methods and are working on the first comprehensive map of mammalian metabolic activity. 

Through this work, we have gained transformative insights into how mammalian metabolism functions. Our studies revealed that the classical metabolic waste product lactate is actually a major circulating fuel. On a molar basis, lactate is the highest flux circulating carbon carrier in mammals. Even in the presence of ample oxygen, glucose is catabolized into circulating lactate, which then serves as a universal energy source for tissues throughout the body.

Most of our metabolic measurements have been made at the level of tissues or tumors. The fundamental biological unit, however, is cells. A major current push in the lab is to enable spatial metabolic measurements at cellular resolution. To this end, we are advancing methods to measure both concentrations and fluxes using MALDI imaging mass spectrometry. These methods are then applied to map metabolism at cellular resolution in tissues and tumors.

How is metabolism regulated?

Metabolic homeostasis has traditionally been studied from a biological perspective, focusing on signaling molecules such as hormones and their effector cascades. This approach has produced major discoveries, including insulin and GLP-1. Yet important gaps remain. Obesity is still widespread and increases the risk of both diabetes and cancer through mechanisms that remain unclear. Reliable personalized dietary guidance remains elusive.

We are revisiting metabolic homeostasis from the perspective of physical chemistry: can the body’s metabolic control systems be quantitatively understood using chemical principles? We find that metabolism is more self-regulating than often appreciated, driven by mass action—where substrate concentrations drive metabolic flux—and by substrate competition, in which metabolites compete for shared enzymes or cofactors. Such competition is widespread, with fat, glucose, lactate, and ketone bodies mutually suppressing each other’s breakdown, a process we term competitive catabolism.

Hormones complement this physicochemical regulation by controlling nutrient storage and release. Glucose stimulates insulin secretion, which suppresses lipolysis—the breakdown of stored fat in adipose tissue into circulating fatty acids. Lactate directly acts on adipose tissue to suppress lipolysis. By lowering circulating fatty acids, such lipolysis suppression accelerates glucose and lactate catabolism.

In obesity, excess fat favors, through mass action, the release of fatty acids into circulation. These fatty acids tend to crowd out glucose metabolism by competing for shared metabolic machinery. As a result, the levels of blood glucose—and therefore insulin—rise. Rising insulin suppresses fat release, but glucose and insulin remain chronically elevated, the hallmark of type 2 diabetes. Thus, the obesity–diabetes connection emerges naturally from the interplay of insulin physiology and chemical mass action. Ongoing work aims to expand our quantitative metabolic models to include a broader range of metabolites and hormones, integrate cancer metabolism, and ultimately enable personalized dietary guidance.