Unlocking the bioactive potential of upcycled antioxidant dietary fibre via machine learning

Reducing food waste and increasing dietary fibre (DF) intake are among the most significant European concerns. Circularity appeared as the answer by creating “upcycled food” from food by-products that can meet DF demands, mitigating resource depletion and minimising the environmental footprint. The development of new DFs and antioxidant dietary fibred (ADFs) from food by-products has been progressed, but knowledge gaps remain concerning the complex relationships among physical structure, physiochemical properties and their potential health effects on the gut. The UPFibre project aims to address these gaps by developing a broad analysis of ADF physical structure, physicochemical properties and their effects on the gut using a multidisciplinary approach enhanced by machine learning (ML) alongside conventional physical, chemical, and biological methodologies. UPFibre intends to find patterns and identify upcycled ADF ingredients’ physical and physicochemical features with gut health benefits. Ultimately, UPFibre aims to establish the first ML-based model to elucidate the relation between ADF structure and gut health effects, advancing the knowledge in this field.