Bitter or not? BitterPredict, a tool for predicting taste from chemical structure

One of the major challenges of masking the bitter taste in medications: an overview of quantitative methods for bitterness

Objective

The aim of the present study was to carry out a systematic research on bitterness quantification to provide a reference for scholars and pharmaceutical developers to carry out drug taste masking research. Significance: The bitterness of medications poses a significant concern for clinicians and patients. Scientifically measuring the intensity of drug bitterness is pivotal for enhancing drug palatability and broadening their clinical utility.

Methods

The current study was carried out by conducting a systematic literature review that identified relevant papers from indexed databases. Numerous studies and research are cited and quoted in this article to summarize the features, strengths, and applicability of quantitative bitterness assessment methods.

Results

In our research, we systematically outlined the classification and key advancements in quantitative research methods for assessing drug bitterness, including in vivo quantification techniques such as traditional human taste panel methods, as well as in vitro quantification methods such as electronic tongue analysis. It focused on the quantitative methods and difficulties of bitterness of natural drugs with complex system characteristics and their difficulties in quantification, and proposes possible future research directions.

Conclusion

The quantitative methods of bitterness were summarized, which laid an important foundation for the construction of a comprehensive bitterness quantification standard system and the formulation of accurate, efficient and rich taste masking strategies.

Molecular insights into binding of bioactive compounds from essential oil of Trachyspermum ammi with human programmed cell death protein 1

The human programmed cell death protein 1 (PD-1) is expressed on the surface of T cells and contributes significantly to tumor immunity. Herein, six major compounds (carvacrol, thymol, β-phellandrene, α-terpinene, myrcene D, and α-pinene) from Trachyspermum ammi were studied for their intermolecular interactions and stability against PD-1. All tested compounds displayed docking energy (-4.2 to -3.7 kcal/mol) with PD-1. The highest docking scores of -4.2 and -4.1 kcal/mol were recorded for carvacrol and thymol, respectively. Also, a 100 ns molecular dynamics simulation predicted the stability of carvacrol- and thymol-docked PD-1 complex. Maximum of < 30 Å and < 12 Å root-mean-square deviation were observed for carvacrol and thymol at the end of the 100 ns simulation with respect to protein (Cα atoms), indicating retention and displacement of carvacrol and thymol from the initial binding pocket, respectively. Moreover, the endpoint binding free energies support the higher binding affinity of carvacrol (-22.87 ± 5.52 kcal/mol) than thymol (-16.83 ± 1.30 kcal/mol). The equicrural states of the respective ligands were supported by the respective root mean square fluctuation, where no significant deviations in the atoms of the ligands were observed. These findings suggest that carvacrol and thymol inhibit the PD-1/PD-L1 axis.Communicated by Ramaswamy H. Sarma.

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.

Comparison of prediction models for soy protein isolate hydrolysates bitterness built using sensory, spectrofluorometric and chromatographic data from varying enzymes and degree of hydrolysis

The bitterness of soy protein isolate hydrolysates prepared using five proteases at varying degree of hydrolysis (DH) and its relation to physicochemical properties, i.e., surface hydrophobicity (H0), relative hydrophobicity (RH), and molecular weight (MW), were studied and developed for predictive modelling using machine learning. Bitter scores were collected from sensory analysis and assigned as the target, while the physicochemical properties were assigned as the features. The modelling involved data pre-processing with local outlier factor; model development with support vector machine, linear regression, adaptive boosting, and K-nearest neighbors algorithms; and performance evaluation by 10-fold stratified cross-validation. The results indicated that alcalase hydrolysates were the most bitter, followed by protamex, flavorzyme, papain, and bromelain. Distinctive correlation results were found among the physicochemical properties, influenced by the disparity of each protease. Among the features, the combination of RH-MW fitted various classification models and resulted in the best prediction performance.

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.

BitterMasS: Predicting Bitterness from Mass Spectra

Bitter compounds are common in nature and among drugs. Previously, machine learning tools were developed to predict bitterness from the chemical structure. However, known structures are estimated to represent only 5-10% of the metabolome, and the rest remain unassigned or "dark". We present BitterMasS, a Random Forest classifier that was trained on 5414 experimental mass spectra of bitter and nonbitter compounds, achieving precision = 0.83 and recall = 0.90 for an internal test set. Next, the model was tested against spectra newly extracted from the literature 106 bitter and nonbitter compounds and for additional spectra measured for 26 compounds. For these external test cases, BitterMasS exhibited 67% precision and 93% recall for the first and 58% accuracy and 99% recall for the second. The spectrum-bitterness prediction strategy was more effective than the spectrum-structure-bitterness prediction strategy and covered more compounds. These encouraging results suggest that BitterMasS can be used to predict bitter compounds in the metabolome without the need for structural assignment of individual molecules. This may enable identification of bitter compounds from metabolomics analyses, for comparing potential bitterness levels obtained by different treatments of samples and for monitoring bitterness changes overtime.

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.

Building a Kokumi Database and Machine Learning-Based Prediction: A Systematic Computational Study on Kokumi Analysis

Kokumi is a subtle sensation characterized by a sense of fullness, continuity, and thickness. Traditional methods of taste discovery and analysis, including those of kokumi, have been labor-intensive and costly, thus necessitating the emergence of computational methods as critical strategies in molecular taste analysis and prediction. In this study, we undertook a comprehensive analysis, prediction, and screening of the kokumi compounds. We categorized 285 kokumi compounds from a previously unreleased kokumi database into five groups based on their molecular characteristics. Moreover, we predicted kokumi/non-kokumi and multi-flavor compositions using six structure-taste relationship models: MLP-E3FP, MLP-PLIF, MLP-RDKFP, SVM-RDKFP, RF-RDKFP, and WeaveGNN feature of Atoms and Bonds. These six predictors exhibited diverse performance levels across two different models. For kokumi/non-kokumi prediction, the WeaveGNN model showed an exceptional predictive AUC value (0.94), outperforming the other models (0.87, 0.90, 0.89, 0.92, and 0.78). For multi-flavor prediction, the MLP-E3FP model demonstrated a higher predictive AUC and MCC value (0.94 and 0.74) than the others (0.73 and 0.33; 0.92 and 0.70; 0.95 and 0.73; 0.94 and 0.64; and 0.88 and 0.69). This data highlights the model's proficiency in accurately predicting kokumi molecules. As a result, we sourced kokumi active compounds through a high-throughput screening of over 100 million molecules, further refined by toxicity and similarity screening. Lastly, we launched a web platform, KokumiPD (https://www.kokumipd.com/), offering a comprehensive kokumi database and online prediction services for users.

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.

Prediction of bitterness based on modular designed graph neural network

Motivation

Bitterness plays a pivotal role in our ability to identify and evade harmful substances in food. As one of the five tastes, it constitutes a critical component of our sensory experiences. However, the reliance on human tasting for discerning flavors presents cost challenges, rendering in silico prediction of bitterness a more practical alternative.

Results

In this study, we introduce the use of Graph Neural Networks (GNNs) in bitterness prediction, superseding traditional machine learning techniques. We developed an advanced model, a Hybrid Graph Neural Network (HGNN), surpassing conventional GNNs according to tests on public datasets. Using HGNN and three other GNNs, we designed BitterGNNs, a bitterness predictor that achieved an AUC value of 0.87 in both external bitter/non-bitter and bitter/sweet evaluations, outperforming the acclaimed RDKFP-MLP predictor with AUC values of 0.86 and 0.85. We further created a bitterness prediction website and database, TastePD (https://www.tastepd.com/). The BitterGNNs predictor, built on GNNs, offers accurate bitterness predictions, enhancing the efficacy of bitterness prediction, aiding advanced food testing methodology development, and deepening our understanding of bitterness origins.

Availability and implementation

TastePD can be available at https://www.tastepd.com, all codes are at https://github.com/heyigacu/BitterGNN.

BitterDB database analysis plus cell stiffness screening identify flufenamic acid as the most potent TAS2R14-based relaxant of airway smooth muscle cells for therapeutic bronchodilation

Rationale: Bitter taste receptors (TAS2Rs) are abundantly expressed in airway smooth muscle cells (ASMCs), which have been recognized as promising targets for bitter agonists to initiate relaxation and thereby prevent excessive airway constriction as the main characteristic of asthma. However, due to the current lack of tested safe and potent agonists functioning at low effective concentrations, there has been no clinically approved TAS2R-based drug for bronchodilation in asthma therapy. This study thus aimed at exploring TAS2R agonists with bronchodilator potential by BitterDB database analysis and cell stiffness screening. Methods: Bitter compounds in the BitterDB database were retrieved and analyzed for their working subtype of TAS2R and effective concentration. Compounds activating TAS2R5, 10, and 14 at < 100 μM effective concentration were identified and subsequently screened by cell stiffness assay using optical magnetic twisting cytometry (OMTC) to identify the most potent to relax ASMCs. Then the compound identified was further characterized for efficacy on various aspects related to relaxation of ASMCs, incl. but not limited to traction force by Fourier transform traction force microscopy (FTTFM), [Ca2+]i signaling by Fluo-4/AM intensity, cell migration by scratch wound healing, mRNA expression by qPCR, and protein expressing by ELISA. The compound identified was also compared to conventional β-agonist (isoproterenol and salbutamol) for efficacy in reducing cell stiffness of cultured ASMCs and airway resistance of ovalbumin-treated mice. Results: BitterDB analysis found 18 compounds activating TAS2R5, 10, and 14 at < 100 μM effective concentration. Cell stiffness screening of these compounds eventually identified flufenamic acid (FFA) as the most potent compound to rapidly reduce cell stiffness at 1 μM. The efficacy of FFA to relax ASMCs in vitro and abrogate airway resistance in vivo was equivalent to that of conventional β-agonists. The FFA-induced effect on ASMCs was mediated by TAS2R14 activation, endoplasmic reticulum Ca2+ release, and large-conductance Ca2+-activated K+ (BKCa) channel opening. FFA also attenuated lipopolysaccharide-induced inflammatory response in cultured ASMCs. Conclusions: FFA as a potent TAS2R14 agonist to relax ASMCs while suppressing cytokine release might be a favorite drug agent for further development of TAS2R-based novel dual functional medication for bronchodilation and anti-inflammation in asthma therapy.

Bitter Phytochemicals as Novel Candidates for Skin Disease Treatment

Skin diseases represent a global healthcare challenge due to their rising incidence and substantial socio-economic burden. While biological, immunological, and targeted therapies have brought a revolution in improving quality of life and survival rates for certain dermatological conditions, there remains a stringent demand for new remedies. Nature has long served as an inspiration for drug development. Recent studies have identified bitter taste receptors (TAS2Rs) in both skin cell lines and human skin. Additionally, bitter natural compounds have shown promising benefits in addressing skin aging, wound healing, inflammatory skin conditions, and even skin cancer. Thus, TAS2Rs may represent a promising target in all these processes. In this review, we summarize evidence supporting the presence of TAS2Rs in the skin and emphasize their potential as drug targets for addressing skin aging, wound healing, inflammatory skin conditions, and skin carcinogenesis. To our knowledge, this is a pioneering work in connecting information on TAS2Rs expression in skin and skin cells with the impact of bitter phytochemicals on various beneficial effects related to skin disorders.

Morus alba L. Leaves (WML) Modulate Sweet (TAS1R) and Bitter (TAS2R) Taste in the Studies on Human Receptors - A New Perspective on the Utilization of White Mulberry Leaves in Food Production?

From the nutritional perspective, the main direction of the utilization of white mulberry (Morus alba L.) parts so far has been to produce dietary supplements or functional foods for individuals with diabetes or over-weight. Its leaves are widely known as a valuable source of bioactive compounds responsible for its antioxidant and antidiabetic effects, both in animals and humans. The authors found that processed leaves can also be investigated as potential bitter and/or sweet taste modulators-an important property of new functional foods. The study aimed to validate the inhibitory effect of Morus alba L. on the TAS2R3 and TAS2R13 bitter taste and TSA1R2/TSA1R3 receptors and determine the changes that the conditioning process caused in such receptors. The effect on the receptors was evaluated in specially transfected HEK293T cells, and the inhibition ratio was measured using the calcium release test. Moreover, the stability of phenolics in the simulated intestinal in vitro digestion process was determined. Results showed that the Morus alba leaf extracts were rich in gallic, chlorogenic and caffeic acids together with rutin and quercetin 3-(6-malonyl)-glucoside, while the conditioning process positively affected their amount. Most identified phenolics were reduced during in vitro digestion. In the taste receptors test, it was found that the phytochemicals from conditioned Morus alba leaf extract enhanced sweet taste, together with a reduction of bitter taste receptor activity in some cases. To conclude, the study has found that Morus alba, especially when conditioned for 4 h, seems to be a valuable modulator of taste, which should be considered in future research as a crucial reason for its new utilization.

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.

Synthesis and evaluation of 2-NMPA derivatives as potential agents for prevention of osteoporosis in vitro and in vivo

Abnormal osteoclast differentiation causes various bone disorders such as osteoporosis. Targeting the formation and activation of osteoclasts has been recognized as an effective approach for preventing osteoporosis. Herein, we synthesized eleven 2-NMPA derivatives which are (2-(2-chlorophenoxy)-N-(4-alkoxy-2-morpholinophenyl) acetamides, and evaluated their suppression effects on osteoclastogenesis in vitro by using TRAP-staining assay. Among the synthesized eleven novel 2-NMPAs, 4-(2-(2-chlorophenoxy)acetamido)-3-morpholinophenyl trifluoromethanesulfonate (11b), 4-(2-(2-chlorophenoxy) acetamido)-3-morpholinophenyl-3-(N-(2-oxo-2-((2-(phenylthio) phenyl) amino) ethyl)methylsulfonamido)benzoate (11d), and 4-(2-(2-chlorophenoxy) acetamido)-3-morpholinophenyl 4-acetamidobenzenesulfonate (11h) displayed highly inhibitory bioactivity on the differentiation of primary osteoclasts. 11h was selected for further investigation of the inhibitory effects and potential mechanism involved in the suppression of osteoclastogenesis. In vitro analysis suggested that 11h inhibited osteoclastogenesis with an IC50 of 358.29 nM, decreased the formation of F-action belts and bone resorption, without interfering cell viability and osteoblast differentiation. Furthermore, the mRNA expressions of osteoclast-specific genes such as Acp5, Nfatc1, Dc-stamp, Atp6v0d2, Mmp9, and Ctsk significantly decreased following 11h treatment. RANKL-induced osteoclast-specific proteins analysis demonstrated that 11h suppressed osteoclast differentiation by downregulating of RANKL-mediated TRAF6 expression, followed by inactivation of PI3K/AKT and IκBα/NF-κB signaling pathways. Finally, 11h inhibited ovariectomy-induced bone loss in vivo. Therefore, the current work highlighted the therapeutic potential of 11h as an anti-osteoporosis lead compound.

Relevance of Phytochemical Taste for Anti-Cancer Activity: A Statistical Inquiry

Targeting inflammation and the pathways linking inflammation with cancer is an innovative therapeutic strategy. Tastants are potential candidates for this approach, since taste receptors display various biological functions, including anti-inflammatory activity (AIA). The present study aims to explore the power different tastes have to predict a phytochemical's anti-cancer properties. It also investigates whether anti-inflammatory phytocompounds also have anti-cancer effects, and whether there are tastes that can better predict a phytochemical's bivalent biological activity. Data from the PlantMolecularTasteDB, containing a total of 1527 phytochemicals, were used. Out of these, only 624 phytocompounds met the inclusion criterion of having 40 hits in a PubMed search, using the name of the phytochemical as the keyword. Among them, 461 phytochemicals were found to possess anti-cancer activity (ACA). The AIA and ACA of phytochemicals were strongly correlated, irrespective of taste/orosensation or chemical class. Bitter taste was positively correlated with ACA, while sweet taste was negatively correlated. Among chemical classes, only flavonoids (which are most frequently bitter) had a positive association with both AIA and ACA, a finding confirming that taste has predictive primacy over chemical class. Therefore, bitter taste receptor agonists and sweet taste receptor antagonists may have a beneficial effect in slowing down the progression of inflammation to cancer.

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.

Paediatric Medicinal Formulation Development: Utilising Human Taste Panels and Incorporating Their Data into Machine Learning Training

This review paper explores the role of human taste panels and artificial neural networks (ANNs) in taste-masking paediatric drug formulations. Given the ethical, practical, and regulatory challenges of employing children, young adults (18-40) can serve as suitable substitutes due to the similarity in their taste sensitivity. Taste panellists need not be experts in sensory evaluation so long as a reference product is used during evaluation; however, they should be screened for bitterness taste detection thresholds. For a more robust evaluation during the developmental phase, considerations of a scoring system and the calculation of an acceptance value may be beneficial in determining the likelihood of recommending a formulation for further development. On the technological front, artificial neural networks (ANNs) can be exploited in taste-masking optimisation of medicinal formulations as they can model complex relationships between variables and enable predictions not possible previously to optimise product profiles. Machine learning classifiers may therefore tackle the challenge of predicting the bitterness intensity of paediatric formulations. While advancements have been made, further work is needed to identify effective taste-masking techniques for specific drug molecules. Continuous refinement of machine learning algorithms, using human panellist acceptability scores, can aid in enhancing paediatric formulation development and overcoming taste-masking challenges.

Metabolome and Transcriptome Integrated Analysis of Mulberry Leaves for Insight into the Formation of Bitter Taste

Mulberry leaves are excellent for health care, confirmed as a 'drug homologous food' by the Ministry of Health, China. The bitter taste of mulberry leaves is one of the main problems that hinders the development of the mulberry food industry. The bitter, unique taste of mulberry leaves is difficult to eliminate by post-processing. In this study, the bitter metabolites in mulberry leaves were identified as flavonoids, phenolic acids, alkaloids, coumarins and L-amino acids by a combined analysis of the metabolome and transcriptome of mulberry leaves. The analysis of the differential metabolites showed that the bitter metabolites were diverse and the sugar metabolites were down-regulated, indicating that the bitter taste of mulberry leaves was a comprehensive reflection of various bitter-related metabolites. Multi-omics analysis showed that the main metabolic pathway related to bitter taste in mulberry leaves was galactose metabolism, indicating that soluble sugar was one of the main factors of bitter taste difference in mulberry leaves. Bitter metabolites play a great role in the medicinal and functional food of mulberry leaves, but the saccharides in mulberry leaves have a great influence on the bitter taste of mulberry. Therefore, we propose to retain bitter metabolites with drug activity in mulberry leaves and increase the content of sugars to improve the bitter taste of mulberry leaves as strategies for mulberry leaf food processing and mulberry breeding for vegetable use.

Bitter Odorants and Odorous Bitters: Toxicity and Human TAS2R Targets

Flavor is perceived through the olfactory, taste, and trigeminal systems, mediated by designated GPCRs and channels. Signal integration occurs mainly in the brain, but some cross-reactivities occur at the receptor level. Here, we predict potential bitterness and taste receptors targets for thousands of odorants. BitterPredict and BitterIntense classifiers suggest that 3-9% of flavor and food odorants have bitter taste, but almost none are intensely bitter. About 14% of bitter molecules are expected to have an odor. Bitterness is more common for unpleasant smells such as fishy, amine, and ammoniacal, while non-bitter odorants often have pleasant smells. Experimental toxicity values suggest that fishy ammoniac smells are more toxic than pleasant smells, regardless of bitterness. TAS2R14 is predicted as the main bitter receptor for odorants, confirmed by in vitro profiling of 10 odorants. The activity of bitter odorants may have implications for physiology due to ectopic expression of taste and smell receptors.

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.

Bitter taste signaling in cancer

Pharmacoresistance of cancer cells to many drugs used in chemotherapy remains a major challenge for the treatment of cancer. Multidrug resistance transporters, especially ATP-binding cassette (ABC) transporters, are a major cause of cancer drug resistance since they translocate a broad range of drug compounds across the cell membrane, extruding them out of the cells. The regulation of ABC transporters by bitter taste receptors (TAS2Rs), which might be activated by specific bitter tasting compounds, was described in several types of cells/organs, becoming a potential target for cancer therapy. TAS2Rs expression has been reported in many organs and several types of cancer, like breast, ovarian, prostate, and colorectal cancers, where their activation was shown to be involved in various biological actions (cell survival, apoptosis, molecular transport, among others). Moreover, many TAS2Rs' ligands, such as flavonoids and alkaloids, with well-recognized beneficial properties, including several anticancer effects, have been reported as potential adjuvants in cancer therapies. In this review, we discuss the potential therapeutic role of TAS2Rs and bitter tasting compounds in different types of cancer as a possible way to circumvent chemoresistance.

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.

The Taste-Masking Mechanism of Chitosan at the Molecular Level on Bitter Drugs of Alkaloids and Flavonoid Glycosides from Traditional Chinese Medicine

Taste masking of traditional Chinese medicines (TCMs) containing multiple bitter components remains an important challenge. In this study, berberine (BER) in alkaloids and phillyrin (PHI) in flavonoid glycosides, which are common bitter components in traditional Chinese medicines, were selected as model drugs. Chitosan (CS) was used to mask their unfriendly taste. Firstly, from the molecular level, we explained the taste-masking mechanism of CS on those two bitter components in detail. Based on those taste-masking mechanisms, the bitter taste of a mixture of BER and PHI was easily masked by CS in this work. The physicochemical characterization results showed the taste-masking compounds formed by CS with BER (named as BER/CS) and PHI (named as PHI/CS) were uneven in appearance. The drug binding efficiency of BER/CS and PHI/CS was 50.15 ± 2.63% and 67.10 ± 2.52%, respectively. The results of DSC, XRD, FTIR and molecular simulation further indicated that CS mainly masks the bitter taste by disturbing the binding site of bitter drugs and bitter receptors in the oral cavity via forming hydrogen bonds between its hydroxyl or amine groups and the nucleophilic groups of BER and PHI. The taste-masking evaluation results by the electronic tongue test confirmed the excellent taste-masking effects on alkaloids, flavonoid glycosides or a mixture of the two kinds of bitter components. The in vitro release as well as in vivo pharmacokinetic results suggested that the taste-masked compounds in this work could achieve rapid drug release in the gastric acid environment and did not influence the in vivo pharmacokinetic results of the drug. The taste-masking method in this work may have potential for the taste masking of traditional Chinese medicine compounds containing multiple bitter components.

Design and Synthesis of Aminopyrimidinyl Pyrazole Analogs as PLK1 Inhibitors Using Hybrid 3D-QSAR and Molecular Docking

Cancer continues to be one of the world's most severe public health issues. Polo-like kinase 1 (PLK1) is one of the most studied members of the polo-like kinase subfamily of serine/threonine protein kinases. PLK1 is a key mitotic regulator responsible for cell cycle processes, such as mitosis initiation, bipolar mitotic spindle formation, centrosome maturation, the metaphase to anaphase transition, and mitotic exit, whose overexpression is often associated with oncogenesis. Moreover, it is also involved in DNA damage response, autophagy, cytokine signaling, and apoptosis. Due to its fundamental role in cell cycle regulation, PLK1 has been linked to various types of cancer onset and progression, such as lung, colon, prostate, ovary, breast cancer, melanoma, and AML. Hence, PLK1 is recognized as a critical therapeutic target in the treatment of various proliferative diseases. PLK1 inhibitors developed in recent years have been researched and studied through clinical trials; however, most of them have failed because of their toxicity and poor therapeutic response. To design more potent and selective PLK1 inhibitors, we performed a receptor-based hybrid 3D-QSAR study of two datasets, possessing similar common scaffolds. The developed hybrid CoMFA (q2 = 0.628, r2 = 0.905) and CoMSIA (q2 = 0.580, r2 = 0.895) models showed admissible statistical results. Comprehensive, molecular docking of one of the most active compounds from the dataset and hybrid 3D-QSAR studies revealed important active site residues of PLK1 and requisite structural characteristics of ligand to design potent PLK1 inhibitors. Based on this information, we have proposed approximately 38 PLK1 inhibitors. The newly designed PLK1 inhibitors showed higher activity (predicted pIC50) than the most active compounds of all the derivatives selected for this study. We selected and synthesized two compounds, which were ultimately found to possess good IC50 values. Our design strategy provides insight into development of potent and selective PLK1 inhibitors.

Structural basis for strychnine activation of human bitter taste receptor TAS2R46

Taste sensing is a sophisticated chemosensory process, and bitter taste perception is mediated by type 2 taste receptors (TAS2Rs), or class T G protein–coupled receptors. Understanding the detailed molecular mechanisms behind taste sensation is hindered by a lack of experimental receptor structures. Here, we report the cryo–electron microscopy structures of human TAS2R46 complexed with chimeric mini–G protein gustducin, in both strychnine-bound and apo forms. Several features of TAS2R46 are disclosed, including distinct receptor structures that compare with known GPCRs, a new “toggle switch,” activation-related motifs, and precoupling with mini–G protein gustducin. Furthermore, the dynamic extracellular and more-static intracellular parts of TAS2R46 suggest possible diverse ligand-recognition and activation processes. This study provides a basis for further exploration of other bitter taste receptors and their therapeutic applications.

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.

Identify Bitter Peptides by Using Deep Representation Learning Features

A bitter taste often identifies hazardous compounds and it is generally avoided by most animals and humans. Bitterness of hydrolyzed proteins is caused by the presence of bitter peptides. To improve palatability, bitter peptides need to be identified experimentally in a time-consuming and expensive process, before they can be removed or degraded. Here, we report the development of a machine learning prediction method, iBitter-DRLF, which is based on a deep learning pre-trained neural network feature extraction method. It uses three sequence embedding techniques, soft symmetric alignment (SSA), unified representation (UniRep), and bidirectional long short-term memory (BiLSTM). These were initially combined into various machine learning algorithms to build several models. After optimization, the combined features of UniRep and BiLSTM were finally selected, and the model was built in combination with a light gradient boosting machine (LGBM). The results showed that the use of deep representation learning greatly improves the ability of the model to identify bitter peptides, achieving accurate prediction based on peptide sequence data alone. By helping to identify bitter peptides, iBitter-DRLF can help research into improving the palatability of peptide therapeutics and dietary supplements in the future. A webserver is available, too.

BitterMatch: recommendation systems for matching molecules with bitter taste receptors

Bitterness is an aversive cue elicited by thousands of chemically diverse compounds. Bitter taste may prevent consumption of foods and jeopardize drug compliance. The G protein-coupled receptors for bitter taste, TAS2Rs, have species-dependent number of subtypes and varying expression levels in extraoral tissues. Molecular recognition by TAS2R subtypes is physiologically important, and presents a challenging case study for ligand-receptor matchmaking. Inspired by hybrid recommendation systems, we developed a new set of similarity features, and created the BitterMatch algorithm that predicts associations of ligands to receptors with ~ 80% precision at ~ 50% recall. Associations for several compounds were tested in-vitro, resulting in 80% precision and 42% recall. The encouraging performance was achieved by including receptor properties and integrating experimentally determined ligand-receptor associations with chemical ligand-to-ligand similarities.

BitterMatch can predict off-targets for bitter drugs, identify novel ligands and guide flavor design. The novel features capture information regarding the molecules and their receptors, which could inform various chemoinformatic tasks. Inclusion of neighbor-informed similarities improves as experimental data mounts, and provides a generalizable framework for molecule-biotarget matching.

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.

Reducing the Bitter Taste of Pharmaceuticals Using Cell-Based Identification of Bitter-Masking Compounds

The palatability of a pharmaceutical preparation is a significant obstacle in developing a patient-friendly dosage form. Bitter taste is an important factor for patients in (i) selecting a certain drug from generic products available in the market and (ii) adhering to a therapeutic regimen. The various methods developed for identification of bitter tasting and bitter-taste modulating compounds present a number of limitations, ranging from limited sensitivity to lack of close correlations with sensory data. In this study, we demonstrate a fluorescence-based assay, analyzing the bitter receptor TAS2R-linked intracellular pH (pH_i) of human gastric parietal (HGT-1) cells as a suitable tool for the identification of bitter tasting and bitter-taste modulating pharmaceutical compounds and preparations, which resembles bitter taste perception. Among the fluorometric protocols established to analyze pH_i changes, one of the most commonly employed assays is based on the use of the pH-sensitive dye SNARF-1 AM. This methodology presents some limitations; over time, the assay shows a relatively low signal amplitude and sensitivity. Here, the SNARF-1 AM methodology was optimized. The identified bicarbonate extrusion mechanisms were partially inhibited, and measurements were carried out in a medium with lower intrinsic fluorescence, with no need for controlling external CO₂ levels. We applied the assay for the screening of flavonoids as potential bitter-masking compounds for guaifenesin, a bitter-tasting antitussive drug. Our findings revealed that eriodictyol, hesperitin and phyllodulcin were the most potent suitable candidates for bitter-masking activity, verified in a human sensory trial.

The druggability of bitter taste receptors for the treatment of neurodegenerative disorders

The delivery of therapeutic drugs to the brain remains a major pharmacology challenge. A complex system of chemical surveillance to protect the brain from endogenous and exogenous toxicants at brain barriers hinders the uptake of many compounds with significant in vitro and ex vivo therapeutic properties. Despite the advances in the field in recent years, the components of this system are not completely understood. Recently, a large group of chemo-sensing receptors, have been identified in the blood-cerebrospinal fluid barrier. Among these chemo-sensing receptors, bitter taste receptors (TAS2R) hold promise as potential drug targets, as many TAS2R bind compounds with recognized neuroprotective activity (quercetin, resveratrol, among others). Whether activation of TAS2R by their ligands contributes to their diverse biological actions described in other cells and tissues is still debatable. In this review, we discuss the potential role of TAS2R gene family as the mediators of the biological activity of their ligands for the treatment of central nervous system disorders and discuss their potential to counteract drug resistance by improving drug delivery to the brain.

Quantification, identification and comparison of oligopeptides on five tea categories with different fermentation degree by Kjeldahl method and ultra-high performance liquid chromatography coupled with quadrupole-orbitrap ultra-high resolution mass spectrometry

Peptides with different lengths or amino sequences could have specific tastes or bio-activities. So far, either the quantity or pattern differences of peptide among various of teas were unknown. Here, firstly, we developed a method for tea oligopeptide quantification and made comparison of their contents. Secondly, we applied ultra-high performance liquid chromatography coupled with quadrupole-orbitrap ultra-high resolution mass spectrometry (UHPLC-Quadrupole-Orbitrap-UHRMS) to sequence oligopeptides. As a result, the total amount of oligopeptides in white tea and dark tea were higher, followed by black tea and green tea, finally with oolong tea. It suggested that withering which undergoes with endogenous protease and post-fermented that undergoes with a participation of exotic micro-organisms were key in oligopeptide enrichment. Thirdly, a total of 902 abundant identified peptides, most of which were tripeptide, tetrapeptide, pentapeptide, and hexapeptide were screened against several existing peptide databases. There were a series of taste peptides and bio-active peptides existing.

Generation of Non-Nucleotide CD73 Inhibitors Using a Molecular Docking and 3D-QSAR Approach

Radiotherapy and chemotherapy are conventional cancer treatments. Around 60% of all patients who are diagnosed with cancer receive radio- or chemotherapy in combination with surgery during their disease. Only a few patients respond to the blockage of immune checkpoints alone, or in combination therapy, because their tumours might not be immunogenic. Under these circumstances, an increasing level of extracellular adenosine via the activation of ecto-5’-nucleotidase (CD73) and consequent adenosine receptor signalling is a typical mechanism that tumours use to evade immune surveillance. CD73 is responsible for the conversion of adenosine monophosphate to adenosine. CD73 is overexpressed in various tumour types. Hence, targetting CD73’s signalling is important for the reversal of adenosine-facilitated immune suppression. In this study, we selected a potent series of the non-nucleotide small molecule inhibitors of CD73. Molecular docking studies were performed in order to examine the binding mode of the inhibitors inside the active site of CD73 and 3D-QSAR was used to study the structure–activity relationship. The obtained CoMFA (q² = 0.844, ONC = 5, r² = 0.947) and CoMSIA (q² = 0.804, ONC = 4, r² = 0.954) models showed reasonable statistical values. The 3D-QSAR contour map analysis revealed useful structural characteristics that were needed to modify non-nucleotide small molecule inhibitors. We used the structural information from the overall docking and 3D-QSAR results to design new, potent CD73 non-nucleotide inhibitors. The newly designed CD73 inhibitors exhibited higher activity (predicted pIC₅₀) than the most active compound of all of the derivatives that were selected for this study. Further experimental studies are needed in order to validate the new CD73 inhibitors.

Prediction of flavor and retention index for compounds in beer depending on molecular structure using a machine learning method

In order to make a preliminary prediction of flavor and retention index (RI) for compounds in beer, this work applied the machine learning method to modeling depending on molecular structure. Towards this goal, the flavor compounds in beer from existing literature were collected. The database was classified into four groups as aromatic, bitter, sulfury, and others. The RI values on a non-polar SE-30 column and a polar Carbowax 20M column from the National Institute of Standards Technology (NIST) were investigated. The structures were converted to molecular descriptors calculated by molecular operating environment (MOE), ChemoPy and Mordred, respectively. By combining the pretreatment of the descriptors, machine learning models, including support vector machine (SVM), random forest (RF) and k-nearest neighbour (kNN) were utilized for beer flavor models. Principal component regression (PCR), random forest regression (RFR) and partial least squares (PLS) regression were employed to predict the RI. The accuracy of the test set was obtained by SVM, RF, and kNN. Among them, the combination of descriptors calculated by Mordred and RF model afforded the highest accuracy of 0.686. R 2 of the optimal regression model achieved 0.96. The results indicated that the models can be used to predict the flavor of a specific compound in beer and its RI value.

Imputation of sensory properties using deep learning

Predicting the sensory properties of compounds is challenging due to the subjective nature of the experimental measurements. This testing relies on a panel of human participants and is therefore also expensive and time-consuming. We describe the application of a state-of-the-art deep learning method, Alchemite™, to the imputation of sparse physicochemical and sensory data and compare the results with conventional quantitative structure-activity relationship methods and a multi-target graph convolutional neural network. The imputation model achieved a substantially higher accuracy of prediction, with improvements in R² between 0.26 and 0.45 over the next best method for each sensory property. We also demonstrate that robust uncertainty estimates generated by the imputation model enable the most accurate predictions to be identified and that imputation also more accurately predicts activity cliffs, where small changes in compound structure result in large changes in sensory properties. In combination, these results demonstrate that the use of imputation, based on data from less expensive, early experiments, enables better selection of compounds for more costly studies, saving experimental time and resources.

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.

Pathogenic genetic variants from highly connected cancer susceptibility genes confer the loss of structural stability

Genetic polymorphisms in DNA damage repair and tumor suppressor genes have been associated with increasing the risk of several types of cancer. Analyses of putative functional single nucleotide polymorphisms (SNP) in such genes can greatly improve human health by guiding choice of therapeutics. In this study, we selected nine genes responsible for various cancer types for gene enrichment analysis and found that BRCA1, ATM, and TP53 were more enriched in connectivity. Therefore, we used different computational algorithms to classify the nonsynonymous SNPs which are deleterious to the structure and/or function of these three proteins. The present study showed that the major pathogenic variants (V1687G and V1736G of BRCA1, I2865T and V2906A of ATM, V216G and L194H of TP53) might have a greater impact on the destabilization of the proteins. To stabilize the high-risk SNPs, we performed mutation site-specific molecular docking analysis and validated using molecular dynamics (MD) simulation and molecular mechanics/Poisson Boltzmann surface area (MM/PBSA) studies. Additionally, SNPs of untranslated regions of these genes affecting miRNA binding were characterized. Hence, this study will assist in developing precision medicines for cancer types related to these polymorphisms.