Computational Gastronomy: The Emerging Data Science of Food, Flavours, and Health

By: Dr Ganesh Bagler*


Computational GastronomyCooking is the core of our cultural identity. Diverse culinary dimensions such as: food; culture; cooking; aroma; taste; and health have been investigated within the scope of gastronomy. The craft of blending ingredients for creating delicious recipes following traditional protocols has hitherto been seen from an artistic perspective. The ability to cook has been arguably central to the evolution of superior human cognitive abilities. Ironically, food is also a key factor behind the lifestyle disorders such as: obesity; cardiovascular disorders; and diabetes. The quest for creating food not only evokes sensory pleasure but also nourishes and enhances health, and has been of paramount importance in gastronomic endeavours.

Computational Gastronomy

The human body is a complex machine which evades understanding of mechanisms involved in the sensory perception of food and health impacts of diet. However, the availability of culinary data and the advent of computational techniques are dramatically changing this situation. Application of data-driven strategies for investigating the gastronomic questions has opened up exciting avenues and also giving rise to a whole new field of ‘Computational Gastronomy’. At the Complex Systems Laboratory, IIIT-Delhi, we have been setting the foundation of this niche interdisciplinary domain by asking questions such as: ‘Why we eat and what we eat?’; ‘What is the molecular basis of taste and odour?’; and ‘How do food ingredients impact health?’. We are attempting to answer these questions braced with the structured data traditional recipes, molecular constituents of ingredients, flavour perceptions of flavour compounds, and health associations of food.

Application of Culinary Fingerprints

Using the data of traditional recipes from across the world, we have devised ‘culinary fingerprints’ that enumerate the patterns in their ingredient composition. Like knowing the law of gravity has enabled us in predicting eclipses and launching satellites. Such fingerprints encapsulate cultural idiosyncrasies and provide means for creating targeted culinary applications. Designing culture-specific food and beverage products; and informed food-beverage pairings, are among some of the applications of culinary fingerprints. The potential space of recipes is practically infinite, because of the presence of diverse ingredients and cooking methods available. Design of algorithms aimed at generating novel recipes following the cuisine-specific culinary traits is another direction that is extremely exciting and holds immense potential to exploit the powers of combinatorial magic.

Structured Repositories

Knowing very well that computational gastronomy is heavily dependent on well-curated and structured data compilations, we have created structured repositories of traditional recipes, flavour profiles of natural ingredients, and health impacts of food.

Flavour DB: Flavour DB (http://cosylab.iiitd.edu.in/flavordb) is a repository of flavour compounds linked to their ingredients, natural source as well as flavour perceptions. Other than facilitating exploration of the flavour profile of ingredients from 34 categories (beverage, cereal, dairy, fruit, herb, meat, spices, vegetable, etc.), it also provides the ‘flavour pairing’ application for identification of ingredients with desirable flavour overlap. Published in the prestigious biomedical journal, Nucleic Acids Research, the database has also been transformed into a mobile application and was launched at the World Heritage Cuisine Festival.

Bitter Sweet: For addressing the imminent epidemic of obesity and diabetes, the lab has been focusing on computational strategies to seek compounds that enhance the perception of sweetness without increasing calories. Towards these efforts, Bitter Sweet (https://cosylab.iiitd.edu.in/bittersweet) is a web server that implements a state-of-the-art machine learning algorithm for the identification of natural as well as synthetic compounds of desirable bitter-sweet taste gradient.

DietRx: Beyond the investigations of taste perception, the lab has also been engaged in addressing the messy picture of food-disease associations, which are often presented with contradictory evidence. By investigating the disease associations of culinary herbs and spices, we have seen that these aromatic ingredients have a broad spectrum of health benefits beyond their antimicrobial utility. Moving beyond the spices, DietRx (https://cosylab.iiitd.edu.in/dietrx) integrates the health impacts of a wide variety of natural ingredients used in cuisines across the world with the information of food chemicals and also links with genetic details. Among the primary use cases of DietRx is the identification of food compounds; those are putatively engaged in regulating health and dietary intervention strategies.

In this century which is witnessing fascinating applications emerging from information sciences, I believe that computational gastronomy will unfold data-driven innovations that will take us closer to develop divergent applications for food, nutrition, and health.


* Assistant Professor, Complex Systems Laboratory, Centre for Computational Biology, IIIT-Delhi, New Delhi

Share Button

Webmaster LBA

Food Marketing & Technology is a monthly magazine published by L.B. Associates Pvt Ltd

Comments are closed