Untargeted mass spectrometry-based metabolomics approach unveils molecular changes in raw and processed foods and beverages

•First large-scale food composition analysis using mass spectral molecular networking.•A methodological pipeline for compound discovery applicable to diverse foods was generated.•Molecular networking and statistical analyses were combined to detect chemical changes in foods.•Brewing, roasting, fermenting, spoiling, and ripening were investigated in example foods.•Time-based chemical changes were detected in yogurt, tea, beef, turkey, and tomato samples. In our daily lives, we consume foods that have been transported, stored, prepared, cooked, or otherwise processed by ourselves or others. Food storage and preparation have drastic effects on the chemical composition of foods. Untargeted mass spectrometry analysis of food samples has the potential to increase our chemical understanding of these processes by detecting a broad spectrum of chemicals. We performed a time-based analysis of the chemical changes in foods during common preparations, such as fermentation, brewing, and ripening, using untargeted mass spectrometry and molecular networking. The data analysis workflow presented implements an approach to study changes in food chemistry that can reveal global alterations in chemical profiles, identify changes in abundance, as well as identify specific chemicals and their transformation products. The data generated in this study are publicly available, enabling the replication and re-analysis of these data in isolation, and serve as a baseline dataset for future investigations.