Giannattasio, Cristina (2024) Development of an optical-biosensor for the detection of volatile organic compounds associated with the traceability of milk samples. [Tesi di dottorato]

	Documento PDF Thesis_Giannattasio_Cristina_PhD_AI-AGR_37thcycle_PARZIALE.pdf Download (780kB)
	Documento PDF Thesis_Giannattasio_Cristina_PhD_AI-AGR_37thcycle_COMPLETO.pdf Visibile a [TBR] Amministratori dell'archivio Download (7MB) | Richiedi una copia

Abstract

Abstract Monitoring food quality, in particular, milk quality, is critical in order to maintain food safety and human health. Therefore, there is an urgent need for the development of fast, sensitive, reliable and cost-effective methods and sensor systems for milk quality monitoring. Cow’s milk is a key nutritional food, making its quality and safety essential. Volatile Organic Compounds (VOCs) are important indicators of milk quality and origin, reflecting factors such as animal metabolism, diet, geographic region, and grazing conditions, which can differentiate milk samples from various farming systems or regions. Recently, the increasing demand for organic and ‘green’ foods is leading to revaluation of extensive farming and grass-fed milk and dairy products. This trend is reviving the research on the relationships between cattle feeding and product flavour. Great attention is being paid to the VOCs, which are closely related to aroma. The milk VOCs are derived from animal metabolism and interaction with the surrounding environment, namely diet and rearing conditions. There is high evidence that the VOCs profile of milk from grazing cattle is different from that of cows fed indoor. My project aims to develop an innovative “impinger” biosensor that utilizes Molecular Recognition Elements (MRE) such as Odorant-Binding Proteins (OBPs) to detect VOCs in milk, to differentiate milk samples from intensive versus extensive farming systems. The OBP-based monitoring biosensor provides the basis for continuous sampling of milk VOCs in air. The impinger capture the VOCs released from milk, which then are collected in a liquid phase. This liquid sample was transferred to the biosensor chamber, where OBPs bind selectively to VOCs. The binding event triggers a Förster Resonance Energy Transfer (FRET) signal that is proportional to the VOCs concentration, allowing the quantification of specific compounds and it is indicative of milk quality. This real-time approach provides a sensitive and cost-effective solution for VOCs monitoring. To verify the binding between OBPs and the VOCs of interest, Head Space Solid-Phase Microextraction coupled with Gas Chromatography-Mass Spectrometry (HS-SPME/GC-MS) was used. This analytical technique provided confirmation of OBP-VOCs interactions and allowed for the characterization of the VOCs profile in milk samples under various conditions. By comparing HS-SPME/GC-MS results with biosensor responses, the binding efficiency and specificity of OBPs for selected VOCs were validated, ensuring reliable detection. To further enhance the biosensor’s sensitivity, a preliminary step was taken towards using artificial intelligence (AI) to predict advantageous mutations in the binding pocket of the pOBP. It is proposed to develop Machine Learning (ML) algorithms to predict which amino acid substitutions would optimize the OBP’s affinity for target VOCs. Different amino acids in the pOBP were systematically mutated and tested for binding performance. Each mutated pOBP variant was evaluated for its sensitivity to VOCs, specifically focusing on compounds most indicative of milk quality changes, such as ketones and aldehydes. This iterative process provided critical insights into the structure-activity relationship of pOBP and allowed for the identification of variants with improved performance, laying the groundwork for future AI-driven optimization.

Downloads

Actions (login required)

Modifica documento