A system for identifying bitter components in honey and application thereof

Abstract

The application provides a honey bitter component identification system and application, and relates to the technical field of honey quality identification.The system comprises a bitter taste identifier and a bitter taste threshold identifier, wherein the bitter taste identifier composed of various fingerprints such as ECFP-546, ECFP-18 and ECFP-149 is used to distinguish the bitter taste and non-bitter taste of honey components, and the bitter taste threshold identifier composed of various structural characteristics such as MaxESateIndex, MaxPartialCharge, PEOE_VSA8 and VSA_EState4 is used to distinguish the minimum bitter taste perception threshold of the components.Through the system, the bitter substances in honey can be quickly and accurately identified, and technical support is provided for the geographical identification, authenticity identification and medicinal function mining of honey.

Description

Technical Field

[0001] The present invention relates to the technical field of honey quality identification, and in particular to a honey bitter component identification system and application. Background Art

[0002] Biological evaluation and biosensor methods are two major approaches for bitter taste identification. Biological evaluation includes sensory evaluation and animal evaluation. The potential toxicity of some bitter components is a major challenge for sensory evaluation, while the main obstacle for animal evaluation is the inability to identify the relationship between animal aversion and perception. The electronic tongue is a commonly used biosensor method for evaluating compound perception. It primarily relies on a multi-sensor array to sense the overall characteristic response signal of a sample, performing simulated identification and quantitative and qualitative analysis. The electronic tongue has the disadvantage that it responds to a wide range of compounds and cannot recognize specific bitter substances with the high specificity of biological receptors. Furthermore, its output is converted into a response value, which does not provide information on the minimum perceptible concentration of a bitter substance. These methods are also costly, time-consuming, and require specialized technical support. They are particularly inadequate when dealing with the complexity of the thousands of compounds in honey. In this context, we propose a system for identifying bitter components in honey. This system consists of a bitterness identifier and a bitterness threshold identifier. This system can directly predict bitter compounds in honey and their threshold values ​​based on the fingerprint and structural characteristics of the compounds. Summary of the Invention

[0003] In response to the shortcomings of existing technologies, the present invention provides a honey bitter component identification system and application, which can quickly and accurately identify bitter substances in honey, providing technical support for the geographical identification, authenticity identification and medicinal function exploration of honey.

[0004] To achieve the above objectives, the technical solution of the present invention is implemented through the following technical solutions:

[0005] A honey bitter component identification system comprises a compound number input module, an InChIKey encoding module, a bitterness recognition module, and a content ranking module; the compound number input module is used to take the names of bitter molecules and non-bitter molecules as input conditions; the InChIKey encoding module is used to convert the names of bitter molecules and non-bitter molecules into InChIKey, then convert the InChIKey into ASCII code, and generate corresponding fingerprint features and structural features; the bitterness recognition module comprises a bitterness identifier and a bitterness threshold identifier, the bitterness identifier discriminates whether honey components are bitter or non-bitter based on a discrimination model constructed based on corresponding fingerprint features; the bitterness threshold identifier is used to discriminate the lowest bitterness perception threshold of a component; and the content ranking module is used to sort the compound content and bitterness threshold to obtain key bitter components in honey.

[0006] Preferably, the discriminant model is composed of a single or combined hydrophobic extended fingerprint of ECFP-4, ECFP-6, ECFP-18, ECFP-114, ECFP-149, ECFP-210, ECFP-308, ECFP-309, ECFP-322, ECFP-499, ECFP-539, ECFP-546, ECFP-691, ECFP-857, ECFP-952, ECFP-1155, ECFP-1668, ECFP-1850, ECFP-2017, or ECFP-2041.

[0007] Preferably, the bitterness threshold identifier is a quantitative model composed of single or combined molecular surface area, charge structure and electronegativity structural parameters such as MaxESateIndex, MaxPartialCharge, PEOE_VSA8, and VSA_EState4.

[0008] The above system is applied to the identification and appraisal of bitter components in food.

[0009] Preferably, the system is applied to identify bitter components in honey.

[0010] Preferably, the system is applied to the identification of bitter components of Quercus balsamifera honey.

[0011] Preferably, the bitter components of Quercus sylvestris honey identified by the system are kynurenic acid, 4-indolecarboxaldehyde, 2,4-dihydroxyquinoline, 8-hydroxyquinoline, indoline, 2,4-dimethylbenzaldehyde, cytidine, ethylmorphine, uniconazole, ketotifen, 7-methylxanthine, DL-3,4-dihydroxyphenyldiol, oleamide, erucamide, 4,2',6'-trihydroxy-4'-methoxy-3',5'-dimethyldihydrochalcone, and difenoconazole.

[0012] The present invention provides a honey bitter component identification system and application, which has the following advantages over the prior art:

[0013] (1) The system of the present invention consists of two parts: a bitterness identifier and a bitterness threshold identifier. The bitterness identifier, which is based on multiple fingerprints such as ECFP-546, ECFP-18, and ECFP-149, is used to distinguish between bitter and non-bitter honey components. The bitterness threshold identifier, which is based on multiple structural characteristics such as MaxESateIndex, MaxPartialCharge, PEOE_VSA8, and VSA_EState4, is used to determine the minimum bitterness perception threshold of the components. The present invention is low-cost, fast, and can identify the bitterness of thousands of compounds in honey without the need for professional technical support.

[0014] (2) The system developed based on the present invention can identify the bitter taste in Quercus truncatula honey and mine a variety of unreported bitter components;

[0015] (3) The present invention provides technical support for the geographical identification, authenticity identification and medicinal functional component mining of specialty honey. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The best bitterness recognition model and the best bitterness threshold model in Example 1 of the present invention;

[0017] Figure 2 This is the key ECFP4 fingerprint of the optimal bitter taste recognition model in Example 1 of the present invention;

[0018] Figure 3 This is the key structural feature of the optimal bitterness threshold model in Example 1 of the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] Example 1:

[0021] Setting up the bitter ingredient identification system:

[0022] We searched the BitterDB open-source database (https: / / bitterdb.agri.huji.ac.il / dbbitter.php) and identified 1,248 bitter molecules and 1,418 non-bitter molecules (including alkaloids, flavonoids, polyphenols, and terpenes). In constructing a bitter identifier, we used Rdkit to read the SMILES of these molecules and calculated the ECFP fingerprint (ECFP4) with a radius of 2 for each molecule. This fingerprint is a 2,048-bitter list of 0s and 1s.

[0023] In constructing a bitterness threshold identifier, a dataset of 153 molecules with a bitterness threshold (such as quinine, tannins, and matrine) was used. These datasets were obtained from the open-source BitterDB database (https: / / bitterdb.agri.huji.ac.il / dbbitter.php). Using molecular properties as descriptors, the model was trained by reducing the number of features to mitigate dimensionality loss and improve generalization. Specifically, the MolecularDescriptorCalculator module in the RDKit package was used to calculate 208 descriptors of the structural properties of all compounds. After removing all features with zero values ​​and high multicollinearity (>0.6), 39 structural features were selected.

[0024] Training and modeling using MaxESateIndex, MaxPartialCharge, PEOE_VSA8, and VSA_EState4: During model training, the dataset was split into a training set and an independent validation set in an 8:2 ratio. The training set was trained using 10-fold cross-validation, which means the model was evenly divided into 10 parts, 9 of which were selected for model training each time, and the remaining part for testing. Each algorithm was optimized using hyperparameters, and after model training, it was tested on the independent validation set.

[0025] The accuracy of the best bitter taste identifier and bitterness threshold identifier is calculated using the confusion matrix, as shown in the following example: Figure 1 As shown. Figure 1 A represents the confusion matrix result of the best bitter taste identifier, with an accuracy of 86% for bitter taste and 83% for non-bitter taste. Figure 1 B represents the accuracy of the best bitterness threshold identifier, and the best model’s goodness of fit (R) on the test set and validation set. 2 ) are both higher than 0.75. These results indicate that both the constructed bitter taste identifier and bitter taste threshold identifier exhibit high accuracy.

[0026] Figure 2 and Figure 3They represent the key features within the optimal bitter taste recognizer and bitter taste threshold recognizer, respectively, which indicate that hydrophobic groups, molecular surface area, charge structure and electronegativity are crucial for bitter taste recognition.

[0027] Example 2:

[0028] The above system was used to identify the bitter components in Quercus truncatula honey:

[0029] The collection time of Quercus truncatum bitter honey was April to May, and the collection years were 2021 and 2023, with a total of 24 samples. 5 g of the Quercus truncatum bitter honey was accurately weighed into a centrifuge tube, and 10 mL of deionized water was added and vortexed until mixed. Subsequently, 10 mL of methanol was added for extraction, and the mixture was centrifuged at 9000 rpm for 30 min. After that, the supernatant was filtered through a 0.22 μm filter membrane to obtain the Quercus truncatum bitter honey component extract.

[0030] 5 mL of methanol and 5 mL of pure water were used to activate the Agilent SPE cartridge, and the above-mentioned bitter oak honey extract was loaded onto the Agilent SPE cartridge. After loading, the cartridge was rinsed with 3 mL of water; the cartridge was further eluted with methanol twice, 4 mL each time, and the eluate was collected and concentrated by nitrogen blow-through, and then reconstituted with 1 mL of methanol; the reconstituted bitter oak honey extract was subjected to non-targeted component identification using ultra-performance liquid chromatography-mass spectrometry equipment to obtain metabolite data of the bitter oak honey extract. The detection method of the ultra-performance liquid chromatography-mass spectrometry equipment is as follows:

[0031] The chromatographic column was an EclipsePlus C18 RRHD (3.0 mm×150 mm, 1.8 μm). The mobile phases were 0.1% formic acid-5 mmol / L ammonium formate in water (C) and 0.1% formic acid-5 mmol / L ammonium formate in methanol (B). The gradient elution program was as follows: 1 min, 99% C; 3 min, 99% C; 15 min, 10% C; 25 min, 1% C; 30 min, 99% C; 35 min, 99% C. The flow rate was 0.25 mL / min, and the injection volume was 0.2 μL.

[0032] Full-scan MS and secondary MS modes were performed. The main mass spectrometry parameters were as follows: sheath gas flow rate was set to 50 Arb; aux gas flow rate was set to 12 Arb; capillary temperature: 330°C; full MS resolution, 70000; AGC target for full MS was 3e6; MS / MS resolution, 17500; AGC target for MS / MS was 1e5; MS / MS isolation window, 4.0 m / z; NCE, 30; spray voltage, 3.5 kV (negative) or 4.0 kV (positive).

[0033] Data processing methods: peak alignment, peak extraction, and normalization were used to obtain a total of 2603 compounds.

[0034] The bitter taste recognition system developed by this patent identified 1,636 bitter substances from 2,603 ​​compounds. These bitter substances mainly come from alkaloids, shikimic acid, phenylpropionic acid and fatty acids.

[0035] A total of 20 high-content bitter substances were obtained through the ranking module, of which 4 bitter substances have been reported, and 16 compounds (kynuric acid, 4-indolecarboxaldehyde, 2,4-dihydroxyquinoline, 8-hydroxyquinoline, indoline, 2,4-dimethylbenzaldehyde, cytidine, ethylmorphine, uniconazole, ketotifen, 7-methylxanthine, DL-3,4-dihydroxyphenyldiol, oleamide, erucamide, 4,2',6'-trihydroxy-4'-methoxy-3',5'-dimethyldihydrochalcone, and difenoconazole) were identified as bitter substances for the first time. The results are shown in Table 1.

[0036] Table 1 Twenty bitter substances with high content in Quercus truncatula honey

[0037]

[0038] In summary, the present invention overcomes the deficiencies of the prior art and can be used to quickly identify the bitter components of honey. In addition to being used for identifying the bitter components of honey, the present invention can also be used for identifying the bitter components of other foods.

[0039] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A system for identifying bitter components in honey, characterized in that: The identification system consists of a compound number input module, an InChIKey encoding module, a bitterness recognition module, and a content ranking module; The compound number input module is used to take the names of bitter molecules and non-bitter molecules as input conditions; The InChIKey encoding module is used to convert the names of bitter molecules and non-bitter molecules into InChIKey, then convert the InChIKey into ASCII code, and generate corresponding fingerprint features and structural features; The bitterness recognition module includes a bitterness identifier and a bitterness threshold identifier. The bitterness identifier is used to distinguish the bitterness and non-bitterness of honey components based on a discrimination model composed of corresponding fingerprint features; the bitterness threshold identifier is used to determine the minimum bitterness perception threshold of the components; The content ranking module is used to sort the compound content and bitterness threshold to obtain the key bitter components in honey; The discriminant model is composed of single or combined hydrophobic extended fingerprints of ECFP-4, ECFP-6, ECFP-18, ECFP-114, ECFP-149, ECFP-210, ECFP-308, ECFP-309, ECFP-322, ECFP-499, ECFP-539, ECFP-546, ECFP-691, ECFP-857, ECFP-952, ECFP-1155, ECFP-1668, ECFP-1850, ECFP-2017, and ECFP-2041; The bitterness threshold identifier is a quantitative model composed of single or combined molecular surface area, charge structure and electronegativity structure parameters of MaxESateIndex, MaxPartialCharge, PEOE_VSA8, and VSA_EState4; The system is applied to the identification of bitter components of Quercus sylvestris nectar, and the bitter components of Quercus sylvestris nectar identified by the system are kynurenic acid, 4-indolecarboxaldehyde, 2,4-dihydroxyquinoline, 8-hydroxyquinoline, indoline, 2,4-dimethylbenzaldehyde, cytidine, ethylmorphine, uniconazole, ketotifen, 7-methylxanthine, DL-3,4-dihydroxyphenyldiol, oleamide, erucamide, 4,2',6'-trihydroxy-4'-methoxy-3',5'-dimethyldihydrochalcone, and difenoconazole.

Images