ChemTastesDB

ChemTastesDB catalogs 2,944 verified and curated organic and inorganic tastants and provides structured molecular and reference data to support research into taste perception.

Key Features:

• Extensive Compound Collection: Contains 2,944 verified and curated tastant compounds categorized into nine classes: five basic tastes (sweet, bitter, umami, sour, salty) and four additional categories (tasteless, non-sweet, multitaste, miscellaneous).
• Detailed Compound Information: Records for each tastant include name, PubChem CID, CAS registry number, canonical SMILES, and taste category.
• Scientific References: Includes references to the scientific sources from which each tastant's data were retrieved.
• Molecular Structure Availability: Provides molecular structures in HyperChem (.hin) format.
• Advanced Analytical Tools: Uses molecular fingerprints to characterize chemical space and supports unsupervised machine learning techniques for exploratory analyses.

Scientific Applications:

• In Silico Taste Prediction: Supports in silico studies to predict taste profiles of unevaluated or yet-to-be-synthesized compounds.
• Molecular–Taste Relationship Analysis: Enables investigation of relationships between molecular structures and taste perceptions using molecular descriptors and classifications.

Methodology:

Molecular fingerprints were computed to characterize chemical space and unsupervised machine learning techniques were applied for exploratory analyses.