INFORM February 2025 Volume 36 (2)
STRUCTURED FATS
inform February 2025, Vol. 36 (2) • 9
And so, the scientific reinvention of edible fats contin ues—aided today by growing data sets and advanced machine learning techniques that seek shortcuts to better molecules. Replacing trans fats in foods is only one goal; developing tasty meat-free and vegan products is another. While market and regulatory trends are paramount, so are the enduring chal lenges of making products that taste as promised, last on shelves, handle temperature changes as designed, and can be made at scale and a profit. Getting these variables consistently right within a given food product can seem as finicky as, well, pie dough. Researchers aim to understand the fundamentals of processes like fat crystallization, which plays a role in shortening and many other foods. “Fat crystallization has always been a complex process,” said Kaustuv Bhattacharya, principal application specialist at International Flavors & Fragrances (IFF) in Braband, Denmark, via email. “In food applications, a variety of fats and oils from vegetable and dairy sources undergo controlled phase tran sitions to provide product structuring.” You introduce even more complexity, he added, when you start to vary the fat
content, aqueous phases, proteins, carbohydrates, emulsifiers, salt, and other aspects of a food formulation. These considerations are all part of the drive to reduce saturated fats, a recent challenge food scientists have been compelled to tackle. In both academia and industry, this area of research has led to a better understanding of the function of cellulosics and liquid oils such as canola, sunflower, and soy bean oil, according to Bhattacharya. “Data science plays a crucial role in this field by analyz ing and extracting predictive insights from various analyti cal instruments,” he said. “By integrating data from different sources, we can develop more accurate models that help us predict and control the crystallization behavior of fats.” This, in turn, allows scientists to enhance the quality and consistency of food products. PREDICTING THE PERFECT FAT Data science and artificial intelligence are helping food scien tists in their quest for alternatives to trans fats and other ingre dients by helping to pinpoint which blends could deliver the desired traits.
Made with FlippingBook Online newsletter creator