by Lorenzo Pallante (Politecnico di Torino), Athanasios Kalogeras (Athena Research Center, Industrial Systems Institute), Marco Agostino Deriu (Politecnico di Torino)
The EU-funded VIRTUOUS project is dedicated to the development of an integrated computational platform working as a virtual tongue able to predict the taste and the organoleptic profile of Mediterranean ingredients. VIRTUOUS integrates mechanistic modelling approaches, machine learning classifiers, algorithms for big data, and cloud computing, all fed by experimental data to predict food taste, knowing its chemical composition.
The Mediterranean diet is often associated with a reduced risk of developing diseases and an increased life expectancy. Several clinical trials highlighted the beneficial effects of the Mediterranean diet in the primary and secondary prevention of cardiovascular diseases, diabetes, and even cancer. In this connection, a rational analysis focused on the chemical compound content of food interacting with specific targets in the human body might be a crucial breakthrough to provide insight into complex biological mechanisms that make the Mediterranean diet result in long-term favourable effects on human homeostasis, which is preserved firstly by taste receptors and driven by nuclear receptors.
In this connection, nature has developed fascinating screening mechanisms to detect healthy or dangerous chemical compounds contained in food. One of the most important control systems for food and drink intake is called “taste”, a sensory modality to evaluate the nutrition content of food and preventing the ingestion of toxic substances. Five basic types of taste are in general accounted: salty, sour, bitter, sweet and umami (by many authors, fat is also considered as a sixth taste). Each taste represents a different nutritional or physiological need.
The common experience is that taste sensasion suddenly arises when food hits our tongue. More in depth, tastants primarily interact with taste-specific proteins, i.e. the taste receptors. The interactions between food chemical compounds and relative taste receptors trigger a subsequent cascade of events involving receptor cells linked to the nervous system, ultimately resulting in the perception of a specific taste. Understanding complex relationships that, from the chemical composition of food, drive towards both its organoleptic profile and its long-term impact on human homeostasis, is a crucial breakthrough oriented to future improvements in applied research targeting health, the market of nutrition supplements, and diet. From the molecular point of view, structural molecular modelling has been recognised as a promising tool for the study of structure-activity relationships and the accurate integration of in silico and experimental methods has provided an up-to-date understanding of the intricate aspects of intermolecular recognition . Moving from structure to function, recently the interest is converging toward promising machine learning, recognised as a powerful approach with yet unexplored capabilities applied to the treatment of large-scale data analysis for efficient extraction of patterns and associations between various types of regions of interest. Several scientific advances have been made in recent years in the field of taste prediction, which should be of paramount importance not only for the food industry but also for the medicine, pharmaceutical and biotechnology sectors 
To face the complexity of the taste-perception process, it is necessary to employ multidisciplinary competencies to develop computational models in a multiscale/multiphysics fashion, moving from chemistry to structure, and from structure to function. This vision is embraced by the EU-supported research actions, namely VIRTUOUS [L1], which aims at developing a user-friendly computational platform able to predict the taste of a food, starting from its chemical composition. VIRTUOUS integrates mechanistic modelling approaches, machine learning classifiers, algorithms for big data, cloud computing, continuously fed and validated by experimental data. (Figure 1).
Figure 1: Schematic representation of the VIRTUOUS platform organisation.
The idea a taste predictor applied to European food products will boost and widen the European food market at a global level, also enhancing Europe's attractiveness as a leading destination. A virtual tongue will allow to “taste” a food even thousands of kilometres away. In this project, the specific application cases will be focused on wine and oil, but the research can be easily extended to any other type of fresh food.
The VIRTUOUS tongue may be also thought to have a “food computer-aided design tool” for EU food technology. For example, based on taste prediction, the VIRTUOUS platform may be used in the future to predict the results of a specific grape graft, and propose what grafting strategy to use to obtain a specific taste or aftertaste. In the same way, all food processing procedures can be optimised and refined through the VIRTUOUS tongue, allowing to obtain different flavours and/or different effects on homeostasis.
The VIRTUOUS research also links taste receptors’ activation to the activity of nuclear receptors, by classifying food molecules based on their ability to interact with specific targets. This aspect will have a great impact on the EU and the global medical field concerning for example immunological disease or cardiovascular pathologies, where a correct diet is commonly complemented by drug therapy. Therefore, the development of a VIRTUOUS tongue has the potential to become a tool to help clinicians in preparing a diet to target specific cell receptors and cell functions and maintain high patient quality of life.
VIRTUOUS embraces a disruptive paradigm concerning modelling physiological and pathological phenomena. The VIRTUOUS vision promises to support a better understanding of the mechanisms underlying the functioning of our physiological and pathological functions and an in-depth understanding of the effects on our bodies of diet, food supplements, drug therapies, or combinations thereof.
All project events and results are also described in the VIRTUOUS official website [L1].
 L. Pallante, M. Malavolta, G. Grasso, et al.: “On the human taste perception: Molecular-level understanding empowered by computational methods”, Trends Food Sci. Technol. 116:445–459, 2021 https://doi.org/10.1016/j.tifs.2021.07.013
 M. Malavolta, L. Pallante, B. Mavkov, et al: “A survey on computational taste predictors”, Eur Food Res. Technol., 2022. https://doi.org/10.1007/s00217-022-04044-5
Marco Agostino Deriu, Politecnico di Torino, Turin, Italy