4357

Computer-Aided Technique for Molecular Taste Recognition, Prediction and Compounds Classifier

Niv Masha, HUJI, Faculty of Agricultural, Food and Environmental Quality Sciences, Biochemistry, Food Science and Nutrition

 

Category

Food & Nutrition, Computer-aided drug discovery  

Keywords

Bitter, Sweet, Machine Learning, Chemical features

Current development stage

TRL7  System prototype demonstration

Application

  • Basic taste qualities like sweet, bitter, sour, salty and umami serve specific functions in identifying food components found in the diet of humans and animals, and are recognized by bitter taste receptors in the oral cavity.
  • It is desirable to identify potential bitter taste of food and pharmaceutical compounds
  • Bitter taste receptors are expressed in extra-oral tissues and are considered as novel therapeutic targets, mainly for asthma.

Our Innovation

A novel and generic computer-aided technique for molecular taste recognition, prediction and compounds classifier to facilitate studying the chemical features associated with bitterness and sweetness Existing BitterDB is a free and searchable database that includes over 680 compounds that were reported to taste bitter to human.

BitterDB predicts taste from chemical structure with ~80% accuracy (BitterPredict) and a novel (unpublished) predictor for intensely bitter compounds

  • Researchers proved a weak relation between bitterness and toxic. Bbitterness is more common in therapeutic drugs than in highly toxic compounds.

Technology

  • A machine learning classifier, BitterPredict predicts whether a compound is bitter or not, based on its chemical structure. The bitterness prediction is based on ligands that fit to a 3D model of receptor or are similar to a particular bitter ligand.
  • Adaptive Boosting based on decision trees machine-learning algorithm applied to the molecules that were represented. Distribution of oral LD50 values of bitter had the same trend as non-bitter.
  • Structure-based and ligand-based prediction of agonists for particular human bitter taste receptors were successfully validated.
  • A structure-based and ligand-based computational approaches to predict novel sweeteners and sweetness enhancers.
  • Sensory tests confirmation.

 4357-1.jpg

Fig. 1: Bitter and non-bitter chemical space

  • A newer classifier predicts whether compound is intensely bitter or not.

Sweetness enhancement:

4357-2.jpg 

Opportunity

  • Bitterness and intense prediction of unknown compounds.
  • Ligands prediction for bitter receptors from different species and rational design of bitterness modulators.
  • New sweeteners and sweetness enhancers with 15-35% less calories for the same sweetness.

Contact for more information:

Amichai Baron
VP, Head of Business Development, Agritech & Envir
+972-8-9489263
Contact ME:
Image CAPTCHA