BitterSweet: Building machine learning models for predicting the bitter and sweet taste of small molecules

Predicting multiple taste sensations with a multiobjective machine learning method

Taste perception plays a pivotal role in guiding nutrient intake and aiding in the avoidance of potentially harmful substances through five basic tastes - sweet, bitter, umami, salty, and sour. Taste perception originates from molecular interactions in the oral cavity between taste receptors and chemical tastants. Hence, the recognition of taste receptors and the subsequent perception of taste heavily rely on the physicochemical properties of food ingredients. In recent years, several advances have been made towards the development of machine learning-based algorithms to classify chemical compounds' tastes using their molecular structures. Despite the great efforts, there remains significant room for improvement in developing multi-class models to predict the entire spectrum of basic tastes. Here, we present a multi-class predictor aimed at distinguishing bitter, sweet, and umami, from other taste sensations. The development of a multi-class taste predictor paves the way for a comprehensive understanding of the chemical attributes associated with each fundamental taste. It also opens the potential for integration into the evolving realm of multi-sensory perception, which encompasses visual, tactile, and olfactory sensations to holistically characterize flavour perception. This concept holds promise for introducing innovative methodologies in the rational design of foods, including pre-determining specific tastes and engineering complementary diets to augment traditional pharmacological treatments.

Computational gastronomy: capturing culinary creativity by making food computable

Cooking, a quintessential creative pursuit, holds profound significance for individuals, communities, and civilizations. Food and cooking transcend mere sensory pleasure to influence nutrition and public health outcomes. Inextricably linked to culinary and cultural heritage, food systems play a pivotal role in sustainability and the survival of life on our planet. Computational Gastronomy is a novel approach for investigating food through a data-driven paradigm. It offers a systematic, rule-based understanding of culinary arts by scrutinizing recipes for taste, nutritional value, health implications, and environmental sustainability. Probing the art of cooking through the lens of computation will open up a new realm of possibilities for culinary creativity. Amidst the ongoing quest for imitating creativity through artificial intelligence, an interesting question would be, 'Can a machine think like a Chef?' Capturing the experience and creativity of a chef in an AI algorithm presents an exciting opportunity for generating a galaxy of hitherto unseen recipes with desirable culinary, flavor, nutrition, health, and carbon footprint profiles.

Machine learning approaches to predict TAS2R receptors for bitterants

Bitter taste involves the detection of diverse chemical compounds by a family of G protein-coupled receptors, known as taste receptor type 2 (TAS2R). It is often linked to toxins and harmful compounds and in particular bitter taste receptors participate in the regulation of glucose homeostasis, modulation of immune and inflammatory responses, and may have implications for various diseases. Human TAS2Rs are characterized by their polymorphism and differ in localization and function. Different receptors can activate various signaling pathways depending on the tissue and the ligand. However, in vitro screening of possible TAS2R ligands is costly and time-consuming. For this reason, in silico methods to predict bitterant-TAS2R interactions could be powerful tools to help in the selection of ligands and targets for experimental studies and improve our knowledge of bitter receptor roles. Machine learning (ML) is a branch of artificial intelligence that applies algorithms to large datasets to learn from patterns and make predictions. In recent years, there has been a record of numerous taste classifiers in literature, especially on bitter/non-bitter or bitter/sweet classification. However, only a few of them exploit ML to predict which TAS2R receptors could be targeted by bitter molecules. Indeed, the shortage and incompleteness of data on receptor-ligand associations in literature make this task non-trivial. In this work, we provide an overview of the state of the art dealing with this specific investigation, focusing on three ML-based models, namely BitterX (2016), BitterSweet (2019) and BitterMatch (2022). This review aims to establish the foundation for future research endeavours focused on addressing the limitations and drawbacks of existing models.

Exploring Bitter and Sweet: The Application of Large Language Models in Molecular Taste Prediction

The perception of bitter and sweet tastes is a crucial aspect of human sensory experience. Concerns over the long-term use of aspartame, a widely used sweetener suspected of carcinogenic risks, highlight the importance of developing new taste modifiers. This study utilizes Large Language Models (LLMs) such as GPT-3.5 and GPT-4 for predicting molecular taste characteristics, with a focus on the bitter-sweet dichotomy. Employing random and scaffold data splitting strategies, GPT-4 demonstrated superior performance, achieving an impressive 86% accuracy under scaffold partitioning. Additionally, ChatGPT was employed to extract specific molecular features associated with bitter and sweet tastes. Utilizing these insights, novel molecular compounds with distinct taste profiles were successfully generated. These compounds were validated for their bitter and sweet properties through molecular docking and molecular dynamics simulation, and their practicality was further confirmed by ADMET toxicity testing and DeepSA synthesis feasibility. This research highlights the potential of LLMs in predicting molecular properties and their implications in health and chemical science.

Bitter flavors and bitter compounds in foods: identification, perception, and reduction techniques

Bitterness is one of the five basic tastes generally considered undesirable. The widespread presence of bitter compounds can negatively affect the palatability of foods. The classification and sensory evaluation of bitter compounds have been the focus in recent research. However, the rigorous identification of bitter tastes and further studies to effectively mask or remove them have not been thoroughly evaluated. The present paper focuses on identification of bitter compounds in foods, structural-based activation of bitter receptors, and strategies to reduce bitter compounds in foods. It also discusses the roles of metabolomics and virtual screening analysis in bitter taste. The identification of bitter compounds has seen greater success through metabolomics with multivariate statistical analysis compared to conventional chromatography, HPLC, LC-MS, and NMR techniques. However, to avoid false positives, sensory recognition should be combined. Bitter perception involves the structural activation of bitter taste receptors (TAS2Rs). Only 25 human TAS2Rs have been identified as responsible for recognizing numerous bitter compounds, showcasing their high structural diversity to bitter agonists. Thus, reducing bitterness can be achieved through several methods. Traditionally, the removal or degradation of bitter substances has been used for debittering, while the masking of bitterness presents a new effective approach to improving food flavor. Future research in food bitterness should focus on identifying unknown bitter compounds in food, elucidating the mechanisms of activation of different receptors, and developing debittering techniques based on the entire food matrix.

Rethinking Sweetener Discovering: Multiparameter Modeling of Molecular Docking Results between the T1R2-T1R3 Receptor and Compounds with Different Tastes

Molecular docking has been widely applied in the discovery of new sweeteners, yet the interpretation of computational results sometimes remains difficult. Here, the interaction between the T1R2-T1R3 sweet taste receptor and 66 tasting compounds, including 26 sweet, 19 bitter, and 21 sour substances was investigated by batch molecular docking processes. Statistical analysis of the docking results generated two novel methods of interpreting taste properties. Quantitative correlation between relative sweetness (RS) and docking results created a multiparameter model to predict sweetness intensity, whose correlation coefficient r = 0.74 is much higher than r = 0.17 for the linear correlation model between sweetness and binding energy. The improved correlation indicated that docking results besides binding energy contain undiscovered information about the ligand-protein interaction. Qualitative discriminant analysis of different tasting molecules generated an uncorrelated linear discriminant analysis (UDLA) model, which achieved an overall 93.1% accuracy in discriminating the taste of molecules, with specific accuracy for verifying sweet, bitter, and sour compounds reaching 88.0%, 92.1%, and 100%. These unprecedented models provide a unique perspective for interpreting computational results and may inspire future research on sweetener discovery.

Predicting and improving complex beer flavor through machine learning

The perception and appreciation of food flavor depends on many interacting chemical compounds and external factors, and therefore proves challenging to understand and predict. Here, we combine extensive chemical and sensory analyses of 250 different beers to train machine learning models that allow predicting flavor and consumer appreciation. For each beer, we measure over 200 chemical properties, perform quantitative descriptive sensory analysis with a trained tasting panel and map data from over 180,000 consumer reviews to train 10 different machine learning models. The best-performing algorithm, Gradient Boosting, yields models that significantly outperform predictions based on conventional statistics and accurately predict complex food features and consumer appreciation from chemical profiles. Model dissection allows identifying specific and unexpected compounds as drivers of beer flavor and appreciation. Adding these compounds results in variants of commercial alcoholic and non-alcoholic beers with improved consumer appreciation. Together, our study reveals how big data and machine learning uncover complex links between food chemistry, flavor and consumer perception, and lays the foundation to develop novel, tailored foods with superior flavors.

Bioinformatics and bioactive peptides from foods: Do they work together?

We live in the Big Data Era which affects many aspects of science, including research on bioactive peptides derived from foods, which during the last few decades have been a focus of interest for scientists. These two issues, i.e., the development of computer technologies and progress in the discovery of novel peptides with health-beneficial properties, are closely interrelated. This Chapter presents the example applications of bioinformatics for studying biopeptides, focusing on main aspects of peptide analysis as the starting point, including: (i) the role of peptide databases; (ii) aspects of bioactivity prediction; (iii) simulation of peptide release from proteins. Bioinformatics can also be used for predicting other features of peptides, including ADMET, QSAR, structure, and taste. To answer the question asked "bioinformatics and bioactive peptides from foods: do they work together?", currently it is almost impossible to find examples of peptide research with no bioinformatics involved. However, theoretical predictions are not equivalent to experimental work and always require critical scrutiny. The aspects of compatibility of in silico and in vitro results are also summarized herein.

Classification of tastants: A deep learning based approach

Predicting the taste of molecules is of critical importance in the food and beverages, flavor, and pharmaceutical industries for the design and screening of new tastants. In this work, we have built deep learning models to classify sweet, bitter, and umami molecules- the three basic tastes whose sensation is mediated by G protein-coupled receptors. An extensive dataset containing 1466 bitter, 1764 sweet, and 238 umami tastants was curated from existing literature. We analyzed the chemical characteristics of the molecules, with special focus on the presence of different functional groups. A deep neural network model based on molecular descriptors and a graph neural network model were trained for taste prediction. The class imbalance due to fewer umami molecules was tackled using special sampling techniques. Both models show comparable performance during evaluation, but the graph-based model can learn task-specific representations from the molecular structure without requiring handcrafted features. We further explain the deep neural network predictions using Shapley additive explanations. Finally, we demonstrated the applicability of the models by screening bitter, sweet, and umami molecules from a large food database. This study develops an in-silico approach to classify molecules based on their taste by leveraging the recent progress in deep learning, which can serve as a powerful tool for tastant design.

A Comprehensive Comparative Analysis of Deep Learning Based Feature Representations for Molecular Taste Prediction

Taste determination in small molecules is critical in food chemistry but traditional experimental methods can be time-consuming. Consequently, computational techniques have emerged as valuable tools for this task. In this study, we explore taste prediction using various molecular feature representations and assess the performance of different machine learning algorithms on a dataset comprising 2601 molecules. The results reveal that GNN-based models outperform other approaches in taste prediction. Moreover, consensus models that combine diverse molecular representations demonstrate improved performance. Among these, the molecular fingerprints + GNN consensus model emerges as the top performer, highlighting the complementary strengths of GNNs and molecular fingerprints. These findings have significant implications for food chemistry research and related fields. By leveraging these computational approaches, taste prediction can be expedited, leading to advancements in understanding the relationship between molecular structure and taste perception in various food components and related compounds.

Classification-based machine learning approaches to predict the taste of molecules: A review

The capacity to discriminate safe from dangerous compounds has played an important role in the evolution of species, including human beings. Highly evolved senses such as taste receptors allow humans to navigate and survive in the environment through information that arrives to the brain through electrical pulses. Specifically, taste receptors provide multiple bits of information about the substances that are introduced orally. These substances could be pleasant or not according to the taste responses that they trigger. Tastes have been classified into basic (sweet, bitter, umami, sour and salty) or non-basic (astringent, chilling, cooling, heating, pungent), while some compounds are considered as multitastes, taste modifiers or tasteless. Classification-based machine learning approaches are useful tools to develop predictive mathematical relationships in such a way as to predict the taste class of new molecules based on their chemical structure. This work reviews the history of multicriteria quantitative structure-taste relationship modelling, starting from the first ligand-based (LB) classifier proposed in 1980 by Lemont B. Kier and concluding with the most recent studies published in 2022.

Data-Driven Elucidation of Flavor Chemistry

Flavor molecules are commonly used in the food industry to enhance product quality and consumer experiences but are associated with potential human health risks, highlighting the need for safer alternatives. To address these health-associated challenges and promote reasonable application, several databases for flavor molecules have been constructed. However, no existing studies have comprehensively summarized these data resources according to quality, focused fields, and potential gaps. Here, we systematically summarized 25 flavor molecule databases published within the last 20 years and revealed that data inaccessibility, untimely updates, and nonstandard flavor descriptions are the main limitations of current studies. We examined the development of computational approaches (e.g., machine learning and molecular simulation) for the identification of novel flavor molecules and discussed their major challenges regarding throughput, model interpretability, and the lack of gold-standard data sets for equitable model evaluation. Additionally, we discussed future strategies for the mining and designing of novel flavor molecules based on multi-omics and artificial intelligence to provide a new foundation for flavor science research.

Potential shared gene signatures and molecular mechanisms between atherosclerosis and depression: Evidence from transcriptome data

BACKGROUND

Atherosclerosis and depression contribute to each other; however, mechanisms linking them at the genetic level remain unexplored. This study aimed to identify shared gene signatures and related pathways between these comorbidities.

METHODS

Atherosclerosis-related datasets were downloaded from the Gene Expression Omnibus database. Differential and weighted gene co-expression network analyses were employed to identify atherosclerosis-related genes. Depression-related genes were downloaded from the DisGeNET database, and the overlaps between atherosclerosis-related genes and depression-related genes were characterized as crosstalk genes. The functional enrichment analysis and protein-protein interaction network were performed in these gene sets. Subsequently, the Boruta algorithm and Recursive Feature Elimination algorithm were performed to identify feature-selection genes. A support vector machine was constructed to measure the accuracy of calculations, and two external validation sets were included to verify the results.

RESULTS

Based on two atherosclerosis-related datasets (GSE28829 and GSE43292), 165 genes were determined as atherosclerosis-related genes. Meanwhile, 1478 depression-related genes were obtained. After intersecting, 24 crosstalk genes were identified, and two pathways, "lipid and atherosclerosis" and "tryptophan metabolism," were revealed as mutual pathways according to the enrichment analysis results. Through the protein-protein interaction network, Molecular Complex Detection plugin, and cytoHubba plugin, PTPRC and MMP9 were identified as the hub gene. Moreover, SLC22A3, CASP1, AMPD3, and PIK3CG were recognized as feature-selection genes. Based on two external validation sets, CASP1 and MMP9 were finally determined as the critical crosstalk genes.

CONCLUSIONS

"Lipid and atherosclerosis" and "tryptophan metabolism" were possibly the pathways of atherosclerosis secondary to depression and depression due to atherosclerosis, respectively. CASP1 and MMP9 were revealed as the most pivotal candidates linking atherosclerosis and depression by mediating these two pathways. Further experimentation is needed to confirm these conclusions.

Machine learning models to predict micronutrient profile in food after processing

The information on nutritional profile of cooked foods is important to both food manufacturers and consumers, and a major challenge to obtaining precise information is the inherent variation in composition across biological samples of any given raw ingredient. The ideal solution would address precision and generability, but the current solutions are limited in their capabilities; analytical methods are too costly to scale, retention-factor based methods are scalable but approximate, and kinetic models are bespoke to a food and nutrient. We provide an alternate solution that predicts the micronutrient profile in cooked food from the raw food composition, and for multiple foods. The prediction model is trained on an existing food composition dataset and has a 31% lower error on average (across all foods, processes and nutrients) than predictions obtained using the baseline method of retention-factors. Our results argue that data scaling and transformation prior to training the models is important to mitigate any yield bias. This study demonstrates the potential of machine learning methods over current solutions, and additionally provides guidance for the future generation of food composition data, specifically for sampling approach, data quality checks, and data representation standards.

Toward a general and interpretable umami taste predictor using a multi-objective machine learning approach

The umami taste is one of the five basic taste modalities normally linked to the protein content in food. The implementation of fast and cost-effective tools for the prediction of the umami taste of a molecule remains extremely interesting to understand the molecular basis of this taste and to effectively rationalise the production and consumption of specific foods and ingredients. However, the only examples of umami predictors available in the literature rely on the amino acid sequence of the analysed peptides, limiting the applicability of the models. In the present study, we developed a novel ML-based algorithm, named VirtuousUmami, able to predict the umami taste of a query compound starting from its SMILES representation, thus opening up the possibility of potentially using such a model on any database through a standard and more general molecular description. Herein, we have tested our model on five databases related to foods or natural compounds. The proposed tool will pave the way toward the rationalisation of the molecular features underlying the umami taste and toward the design of specific peptide-inspired compounds with specific taste properties.

Development and assessment of novel machine learning models to predict medication non-adherence risks in type 2 diabetics

Background

Medication adherence is the main determinant of effective management of type 2 diabetes, yet there is no gold standard method available to screen patients with high-risk non-adherence. Developing machine learning models to predict high-risk non-adherence in patients with T2D could optimize management.

Methods

This cross-sectional study was carried out on patients with T2D at the Sichuan Provincial People's Hospital from April 2018 to December 2019 who were examined for HbA1c on the day of the survey. Demographic and clinical characteristics were extracted from the questionnaire and electronic medical records. The sample was randomly divided into a training dataset and a test dataset with a radio of 8:2 after data preprocessing. Four imputing methods, five sampling methods, three screening methods, and 18 machine learning algorithms were used to groom data and develop and validate models. Bootstrapping was performed to generate the validation set for external validation and univariate analysis. Models were compared on the basis of predictive performance metrics. Finally, we validated the sample size on the best model.

Results

This study included 980 patients with T2D, of whom 184 (18.8%) were defined as medication non-adherence. The results indicated that the model used modified random forest as the imputation method, random under sampler as the sampling method, Boruta as the feature screening method and the ensemble algorithms and had the best performance. The area under the receiver operating characteristic curve (AUC), F1 score, and area under the precision-recall curve (AUPRC) of the best model, among a total of 1,080 trained models, were 0.8369, 0.7912, and 0.9574, respectively. Age, present fasting blood glucose (FBG) values, present HbA1c values, present random blood glucose (RBG) values, and body mass index (BMI) were the most significant contributors associated with risks of medication adherence.

Conclusion

We found that machine learning methods could be used to predict the risk of non-adherence in patients with T2D. The proposed model was well performed to identify patients with T2D with non-adherence and could help improve individualized T2D management.

Taste GPCRs and their ligands

Taste GPCRs are expressed in taste buds on the tongue and play a key role in food choice and consumption. They are also expressed extra-orally, with various physiological roles that are currently under study. Unraveling the roles of these receptors relies on the knowledge of their ligands. Combining sensory, cell-based and computational approaches enabled the discovery of numerous agonists and several antagonists. Here we provide a short overview of taste receptor families, main recent methods for ligands discovery, and current sources of information about known ligands. The future directions that are likely to impact the taste GPCR field include focus on ligand interactions with naturally occurring polymorphisms, as well as harnessing the power of CryoEM and of multiple signaling readout techniques.

BoostSweet: Learning molecular perceptual representations of sweeteners

The development of safe artificial sweeteners has attracted considerable interest in the food industry. Previous machine learning (ML) studies based on quantitative structure-activity relationships have provided some molecular principles for predicting sweetness, but these models can be improved via the chemical recognition of sweetness active factors. Our ML model, a soft-vote ensemble model that has a light gradient boosting machine and uses both layered fingerprints and alvaDesc molecular descriptor features, demonstrates state-of-the-art performance, with an AUROC score of 0.961. Based on an analysis of feature importance and dataset, we identified that the number of nitrogen atoms that serve as hydrogen bond donors in molecules can play an essential role in determining sweetness. These results potentially provide an advanced understanding of the relationship between molecular structure and sweetness, which can be used to design new sweeteners based on molecular structural dependence.

Identification of novel umami molecules via QSAR models and molecular docking

Umami substances can increase the overall taste of food and bring pleasure to people. However, it is still challenging to identify the umami molecules through virtual screening due to the crystal structure of the umami receptor being undefined. Herein, based on the hypothesis that the molecules with bitter and sweet taste characteristics may be umami molecules, this study proposed an in silico method to identify novel umami-tasting molecules in batch from SWEET-DB and BitterDB databases via the QSAR models, PCA, molecular docking and electronic tongue analysis. In total, 169 potential umami molecules were identified through QSAR modeling, PCA, and molecular docking. Of the 169 molecules, 18 were randomly selected, and all were identified as umami molecules via electronic tongue analysis. Among the 18 chosen molecules, 10 molecules could be traced back to their concentration range in food, and finally, 8 molecules were predicted to be nontoxic. This work provides a simple and efficient strategy to identify novel umami molecules, holding an excellent promise for demonstrating the crystal structure of umami receptors and taste-sensing mechanisms. Furthermore, this study opens the possibility for the practical application of new umami molecules in food.

ChemTastesDB: A curated database of molecular tastants

•

A chemical database called ChemTastesDB was created, which consists of 2944 organic and inorganic tastants.

•

Tastants have been curated and classified into 5 basic and 4 non-basic taste categories.

•

The chemical space of the ChemTastesDB has been analyzed with unsupervised machine learning.

•

ChemTastesDB is freely available on line and provides support for decision-making for designing new tastants.

The purpose of this work is the creation of a chemical database named ChemTastesDB that includes both organic and inorganic tastants. The creation, curation pipeline and the main features of the database are described in detail. The database includes 2944 verified and curated compounds divided into nine classes, which comprise the five basic tastes (sweet, bitter, umami sour and salty) along with four additional categories: tasteless, non-sweet, multitaste and miscellaneous. ChemTastesDB provides the following information for each tastant: name, PubChem CID, CAS registry number, canonical SMILES, class taste and references to the scientific sources from which data were retrieved. The molecular structure in the HyperChem (.hin) format of each chemical is also made available. In addition, molecular fingerprints were used for characterizing and analyzing the chemical space of tastants by means of unsupervised machine learning. ChemTastesDB constitutes a useful tool to the scientific community to expand the information of taste molecules and to assist in silico studies for the taste prediction of unevaluated and as yet unsynthetized compounds, as well as the analysis of the relationships between molecular structure and taste. The database is freely accessible at https://doi.org/10.5281/zenodo.5747393.

A survey on computational taste predictors

Taste is a sensory modality crucial for nutrition and survival, since it allows the discrimination between healthy foods and toxic substances thanks to five tastes, i.e., sweet, bitter, umami, salty, and sour, associated with distinct nutritional or physiological needs. Today, taste prediction plays a key role in several fields, e.g., medical, industrial, or pharmaceutical, but the complexity of the taste perception process, its multidisciplinary nature, and the high number of potentially relevant players and features at the basis of the taste sensation make taste prediction a very complex task. In this context, the emerging capabilities of machine learning have provided fruitful insights in this field of research, allowing to consider and integrate a very large number of variables and identifying hidden correlations underlying the perception of a particular taste. This review aims at summarizing the latest advances in taste prediction, analyzing available food-related databases and taste prediction tools developed in recent years.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00217-022-04044-5.

A novel multi-layer prediction approach for sweetness evaluation based on systematic machine learning modeling

Nowadays, computational approaches have drawn more and more attention when exploring the relationship between sweetness and chemical structure instead of traditional experimental tests. In this work, we proposed a novel multi-layer sweetness evaluation system based on machine learning methods. It can be used to evaluate sweet properties of compounds with different chemical spaces and categories, including natural, artificial, carbohydrate, non-carbohydrate, nutritive and non-nutritive ones, suitable for different application scenarios. Furthermore, it provided quantitative predictions of sweetness. In addition, sweetness-related chemical basis and structure transforming rules were obtained by using molecular cloud and matched molecular pair analysis (MMPA) methods. This work systematically improved the data quality, explored the best machine learning algorithm and molecular characterizing strategy, and finally obtained robust models to establish a multi-layer prediction system (available at: https://github.com/ifyoungnet/ChemSweet). We hope that this study could facilitate food scientists with efficient screening and precise development of high-quality sweeteners.

CBDPS 1.0: A Python GUI Application for Machine Learning Models to Predict Bitter-Tasting Children's Oral Medicines

Bitter tastes are innately aversive and are thought to help protect animals from consuming poisons. Children are extremely sensitive to drug tastes, and their compliance is especially poor with bitter medicine. Therefore, judging whether a drug is bitter and adopting flavor correction and taste-masking strategies are key to solving the problem of drug compliance in children. Although various machine learning models for bitterness and sweetness prediction have been reported in the literature, no learning model or bitterness database for children's medication has yet been reported. In this study, we trained four different machine learning models to predict bitterness. The goal of this study was to develop and validate a machine learning model called the "Children's Bitter Drug Prediction System" (CBDPS) based on Tkinter, which predicts the bitterness of a medicine based on its chemical structure. Users can enter the Simplified Molecular-Input Line-Entry System (SMILES) formula for a single compound or multiple compounds, and CBDPS will predict the bitterness of children's medicines made from those XGBoost-Molecular ACCess System (XgBoost-MACCS) model yielded an accuracy of 88% under cross-validation.

Bitter taste in silico: A review on virtual ligand screening and characterization methods for TAS2R-bitterant interactions

The growing pharmaceutical interest in the human bitter taste receptors (hTAS2Rs) has two dimensions; i) evaluation of the bitterness of active pharmaceutical compounds, in order to develop strategies for improving patients' adherence to medication, and ii) application of ligands for extra-cellular hTAS2Rs for potential preventive therapeutic achievements. The result is an increasing demand on robust tools for bitterness assessment and screening the receptor-ligand affinity. In silico tools are useful for aiding experimental-screening, as well as to elucide ligand-receptor interactions. In this review, the ligand-based and structure-based approaches are described as the two main in silico tools for bitter taste analysis. The strengths and weaknesses of each approach are discussed. Both approaches provide key tools for understanding and exploiting bitter taste for human health applications.

Intense bitterness of molecules: Machine learning for expediting drug discovery

Drug development is a long, expensive and multistage process geared to achieving safe drugs with high efficacy. A crucial prerequisite for completing the medication regimen for oral drugs, particularly for pediatric and geriatric populations, is achieving taste that does not hinder compliance. Currently, the aversive taste of drugs is tested in late stages of clinical trials. This can result in the need to reformulate, potentially resulting in the use of more animals for additional toxicity trials, increased financial costs and a delay in release to the market. Here we present BitterIntense, a machine learning tool that classifies molecules into "very bitter" or "not very bitter", based on their chemical structure. The model, trained on chemically diverse compounds, has above 80% accuracy on several test sets. Our results suggest that about 25% of drugs are predicted to be very bitter, with even higher prevalence (~40%) in COVID19 drug candidates and in microbial natural products. Only ~10% of toxic molecules are predicted to be intensely bitter, and it is also suggested that intense bitterness does not correlate with hepatotoxicity of drugs. However, very bitter compounds may be more cardiotoxic than not very bitter compounds, possessing significantly lower QPlogHERG values. BitterIntense allows quick and easy prediction of strong bitterness of compounds of interest for food, pharma and biotechnology industries. We estimate that implementation of BitterIntense or similar tools early in drug discovery process may lead to reduction in delays, in animal use and in overall financial burden.

RecipeDB: a resource for exploring recipes

Cooking is the act of turning nature into the culture, which has enabled the advent of the omnivorous human diet. The cultural wisdom of processing raw ingredients into delicious dishes is embodied in their cuisines. Recipes thus are the cultural capsules that encode elaborate cooking protocols for evoking sensory satiation as well as providing nourishment. As we stand on the verge of an epidemic of diet-linked disorders, it is eminently important to investigate the culinary correlates of recipes to probe their association with sensory responses as well as consequences for nutrition and health. RecipeDB (https://cosylab.iiitd.edu.in/recipedb) is a structured compilation of recipes, ingredients and nutrition profiles interlinked with flavor profiles and health associations. The repertoire comprises of meticulous integration of 118 171 recipes from cuisines across the globe (6 continents, 26 geocultural regions and 74 countries), cooked using 268 processes (heat, cook, boil, simmer, bake, etc.), by blending over 20 262 diverse ingredients, which are further linked to their flavor molecules (FlavorDB), nutritional profiles (US Department of Agriculture) and empirical records of disease associations obtained from MEDLINE (DietRx). This resource is aimed at facilitating scientific explorations of the culinary space (recipe, ingredient, cooking processes/techniques, dietary styles, etc.) linked to taste (flavor profile) and health (nutrition and disease associations) attributes seeking for divergent applications. Database URL:  https://cosylab.iiitd.edu.in/recipedb.

In-silico screening of database for finding potential sweet molecules: A combined data and structure based modeling approach

In this study, we present a framework comprises of several independent modules which are built upon data based (structure activity relationship and classification model) and structure (molecular docking) based for identifying possible sweeteners from a vast database of natural molecules. A large database, Universal Natural Products Database (UNPD) consisting of 213,210 compounds was screened using the developed framework. At first, 10,184 molecules structurally similar to the known sweeteners were identified in the database. Further, 1924 molecules from these screened molecules were classified as sweet molecules. The shortlisted 1354 molecules were subjected to ADMET analysis. Finally, 60 molecules were arrived at with no toxicity and acceptable oral bioavailability as potential sweetener candidates. Further, molecular docking of these molecules on sweet taste receptor performed to obtain their binding energy, binding sites and correlation with sweetness index. The developed framework offers a convenient route for fast screening of molecules prior to synthesis and testing.

Novel scaffold of natural compound eliciting sweet taste revealed by machine learning

Sugar replacement is still an active issue in the food industry. The use of structure-taste relationships remains one of the most rational strategy to expand the chemical space associated to sweet taste. A new machine learning model has been setup based on an update of the SweetenersDB and on open-source molecular features. It has been implemented on a freely accessible webserver. Cellular functional assays show that the sweet taste receptor is activated in vitro by a new scaffold of natural compounds identified by the in silico protocol. The newly identified sweetener belongs to the lignan chemical family and opens a new chemical space to explore.

Toward the Identification of Extra-Oral TAS2R Agonists as Drug Agents for Muscle Relaxation Therapies via Bioinformatics-Aided Screening of Bitter Compounds in Traditional Chinese Medicine

Significant advances have been made in the past decade in mapping the distributions and the physiological functions of extra-oral bitter taste receptors (TAS2Rs) in non-gustatory tissues. In particular, it has been found that TAS2Rs are expressed in various muscle tissues and activation of TAS2Rs can lead to muscle cell relaxation, which suggests that TAS2Rs may be important new targets in muscle relaxation therapy for various muscle-related diseases. So far, however, there is a lack of potent extra-oral TAS2R agonists that can be used as novel drug agents in muscle relaxation therapies. Interestingly, traditional Chinese medicine (TCM) often characterizes a drug’s property in terms of five distinct flavors (bitter, sweet, sour, salty, and pungent) according to its taste and function, and commonly regards “bitterness” as an intrinsic property of “good medicine.” In addition, many bitter flavored TCM are known in practice to cause muscle relaxation after long term use, and in lab experiments the compounds identified from some bitter flavored TCM do activate TAS2Rs and thus relax muscle cells. Therefore, it is highly possible to discover very useful extra-oral TAS2R agonists for muscle relaxation therapies among the abundant bitter compounds used in bitter flavored TCM. With this perspective, we reviewed in literature the distribution of TAS2Rs in different muscle systems with a focus on the map of bitter flavored TCM which can regulate muscle contractility and related functional chemical components. We also reviewed the recently established databases of TCM chemical components and the bioinformatics software which can be used for high-throughput screening and data mining of the chemical components associated with bitter flavored TCM. All together, we aim to present a knowledge-based approach and technological platform for identification or discovery of extra-oral TAS2R agonists that can be used as novel drug agents for muscle relaxation therapies through screening and evaluation of chemical compounds used in bitter flavored TCM.