INFORM February 2025 Volume 36 (2)

8 • inform February 2025, Vol. 36 (2)

Data science reveals

new possibilities in food chemistry

Christina Nunez

Many bakers swear by using shortening for a delicious pie crust, and with good reason: Mixed in with flour, the fat forms layers that promote flakiness and inhibit the formation of gluten chains that can make a dough tough or chewy. Added to frosting, shortening lends structure and fluffi ness. You can fry with it; you can grease a pan with it. This versatility has kept commercial shorten ing in kitchen pantries for over a century.

If shortening is an ideal fat for baking, it is also a prime example of how the definition of a ‘model’ food evolves over time—usually in response to one market force or another. The shortening you find on grocery store shelves today is a distant relative to the earliest ver sions, having evolved from partially hydrogenated vegetable oil to a blend of other oil types that leave dreaded trans fats behind. Here is where another kind of model comes in with modern fats: one created with computer data. As the food industry works to keep in step with changes in supply chains, health science, and fla vor trends, data science and artificial intelligence (AI) are increasingly essential tools. While the methods have become more sophisticated over the years, the goals have not changed: We want fats to carry flavor and deliver texture without breaking the bank or posing major health risks. SEEKING ALTERNATIVES TO ‘THE WORST’ FAT In the early 1900s, Crisco billed itself as a ‘digestible’ alternative to lard in cooking. Derived from cottonseed oil—and later, palm, soy bean, and other oil sources—Crisco and other commercial shorten ings were affordable, shelf-stable, and had a neutral flavor. Plus, they were plant-based, lending to a perception of healthfulness compared to animal fats. By the 1990s, however, what was once seen as an innocuous modern convenience became a public health enemy. The hydrogena tion process that turned oils into solids at room temperature created trans fats, which studies linked to heart disease. Artificial trans fats became “the worst type of fat to eat,” as a post at the Mayo Clinic’s website puts it, and the US Food and Drug Administration banned them in 2018.

• The food industry seeks fats that can deliver texture, flavor, and a long shelf life. • Health concerns over trans and saturated fats have prompted a search for alternatives. • Food scientists are turning to data analysis and artificial intelligence as a way of pinpointing potential new ingredients. • Exploring the characteristics and molecular structure of different triglycerides through computer models can be faster and cheaper than performing physical experiments and sensory panels.