InSyBio announces Novel of Multiple Taste Predictor based on a Multiobjective Machine Learning method

A novel predictor for multiple taste sensations based on a multi-objective Machine Learning method was published by InSyBio in the NPJ Science of Food journal of the Springer Nature publishing company (https://rdcu.be/dOVbA). This new predictor, namely Virtuous Multi-Taste, was developed in collaboration with a team of exceptional researchers from Politecnico di Torino (Italy), Industrial Systems Institute (Greece) and Dalle Molle Institute for Artificial Intelligence (Switzerland) in the framework of VIRTUOUS Project, an EU Horizon 2020 program.

This work provides a valuable tool for predicting the taste of chemical compounds based on their most important physiochemical properties. The tool is designed to analyze different compounds and provide as output the percentage of contribution of four different tastes- sweet, umami, bitter and other- to the overall taste of each query compound. The multi-class predictor was created based on InSyBio’s multi-objective evolutionary optimization algorithm that is implemented in InSyBio Suite, showcasing the wide range of applications of InSyBio’s biomarker discovery and predictive analytic algorithms and pipelines.

The Virtuous Multi-Taste tool will accelerate taste perception research by providing nutrition and pharmaceutical companies and laboratories with a much-needed tool for compound design based on taste. The tool itself can be accessed through a web-based interface: https://virtuous.isi.gr/#/virtuous-multitaste.

Image: (a) The Virtuous Multi-Taste web-based interface and the result of compound search indicating (b) each compound’s structure and (c) the percentage of each taste’s representation in the final compound taste.

The vision of the Virtuous project, as previously stated by Dr. Marco Agostino Deriu, was to “build an intelligent machine able to decrypt a sensation by modelling the biophysical architecture of our body, covering all required competencies for this ambitious challenge, starting from molecular and multiscale modelling techniques, to artificial intelligence algorithms; experimental facilities to analyze natural product typical of the Mediterranean diet such as wine and oil“

Labros Digonis, CEO of InSyBio, commented on the importance of the tool:

We are proud of acquiring, at last, the ability to predict the taste of chemical compounds based on their physiochemical properties. The potential to revolutionize multiple industries by improving product development processes, enhancing consumer satisfaction and fostering innovation in taste-related research is inherent to this revolutionary discovery and is here to stay. We will be seeing a lot of new, cost-efficient, more targeted and senses-satisfying products based on this. Congratulations to the pioneer team!”

About Virtuous: The EU-funded VIRTUOUS project (Marie Skłodowska-Curie Research and Innovation Staff Exchange (RISE) – https://virtuoush2020.com/) is dedicated to the development of a virtual tongue made by an integrated computational framework that can screen food for natural components by binding taste receptors. This international consortium includes eight academic and industry participants from four European countries. Using a variety of computational and experimental data, partners plan to create an algorithm for the prediction of the organoleptic profile of the food based on its chemical composition. The potential application of this platform as a taste predictor is in the food industry, modern precision agriculture, and as a link to neuroscience for taste determination.

About InSyBio: InSyBio is an international pioneer bioinformatics revolutionizing the medical field, clinical trials and food/nutrition industries through targeted biomarker discovery and highly-accurate predictive analytics.

InSyBio solutions provide the R&D departments of Pharma, Research Institutes and Food/Nutrition companies, with the means both analytically and computationally, to meet their challenging research and innovation goals (such as robust biomarker discovery and accurate predictive models). This is done through a sophisticated, software-as-a-service, machine learning platform which comprehensively integrates multi-omics data reducing the number of samples and additional validation experiments needed. Since 2013, InSyBio has worked with hundreds of academics, large and small companies, hospitals and research centers allowing them to succeed in their biomarker discovery tasks with an increase of at least 10% in predictive accuracy and decrease of at least 80% in overall time and cost.

Scroll to top