Data-driven and knowledge-based analytical valorization of natural products for sustainable and innovative applications

Analytical chemistry plays a key role in the valorization of natural products. When combined with sustainable technologies such as green biorefineries, the stakes are even higher. Valorization has a broad meaning. For this reason, this work is divided into five different chapters. Three different levels of valorization are discussed. The introduction (chapter one) begins with a description of the biorefinery process, focusing on Medicago sativa. Then the analytical tools, the data analysis strategies, and their combination are discussed. Chapter two is a description of the proteomic and phenolic characterization of biorefinery products. The valorization starts from here. Design of Experiments (DoE) was used to identify the best combinations of factors to optimize the extraction procedures. These were then evaluated and modified according to their biocompatibility with cells. The profile of proteins, protein hydrolysates, and metabolites was defined using liquid chromatography with tandem mass spectrometry. The second level is the investigation of the biological effects of complex natural mixtures. Specifically, the pro-apoptotic activity was studied on colorectal cancer cells as an epithelial proliferation model. Data-driven and knowledge-based approaches were used to generate hypotheses about their mechanisms of action. Chapter three is divided into two sections. The first focuses on Medicago sativa biorefinery products. The second focuses on protein extracts from four quinoa varieties. Overall, the basic idea behind this chapter was to demonstrate the combination of data modeling and critical thinking as a powerful tool for transforming raw data into understanding, new perspectives, and knowledge. The third level is about the optimization of the biorefinery process. In the fourth chapter, near-infrared spectroscopy was used to develop regression and classification models for process monitoring. Despite the satisfactory results, this chapter aims to highlight the tension between optimization and generalization. Especially when it comes to small datasets (typical of laboratory and pilot scale). Finally, chapter five provides a summary of the results and future perspectives. In conclusion, this dissertation aims to present different valorization strategies. But at the same time, all the challenges that can arise and a way to deal with them. Especially using the different paradigms of data analysis.