Odor molecule and olfactory receptor reaction prediction method and device based on multi-mode deep learning network

Abstract

The invention relates to an odor molecule and olfactory receptor reaction prediction method and device based on a multi-mode deep learning network, and the method comprises the steps: firstly, extracting physical and chemical descriptors of odor molecules as the characteristics of the physical and chemical descriptors through the interaction data between the odor molecules and olfactory receptors, and employing a deep learning network model to predict the reaction between the odor molecules and the olfactory receptors; the method comprises the following steps: firstly, judging whether an unknown odor molecule can excite an olfactory receptor or not, secondly, judging the possibility of reaction between the odor molecule and each olfactory receptor through a network calculation model by adopting odor molecule sequence data, and then, further introducing odor molecule structure image data, and calculating the reaction probability between the odor molecule and each olfactory receptor. And finally, in order to improve the accuracy and reliability of a recognition result, determining a final reaction olfactory receptor of the odorant molecule through a calculation result obtained through analysis, namely two groups of probability distributions. According to the method, the recognition accuracy of the odor molecules and the olfactory receptors is remarkably improved, meanwhile, the dependence on an experimental method is reduced, and a new tool is provided for chemical and biological research.

Description

Technical Field

[0001] The present invention belongs to the technical field of prediction of odor molecules and olfactory receptor reactions, and specifically relates to a method and device for predicting the reactions of odor molecules and olfactory receptors based on a multimodal deep learning network. Background Art

[0002] Analyzing the interaction between odor molecules and olfactory receptors is crucial for fields such as drug design and flavor development. In the field of drug design, accurately predicting the interaction between odor molecules and olfactory receptors is crucial for discovering new drugs and optimizing lead compounds; in the flavor industry, understanding and predicting how odor molecules interact with olfactory receptors can accelerate the discovery and development of new flavors; in addition, the field of environmental science also pays more attention to the impact of chemicals on biological olfactory receptor systems, which is of great significance for assessing environmental risks and formulating safety standards.

[0003] In chemical and biological research, analyzing odor molecules and the olfactory receptors they react to is of great significance for understanding olfactory perception, as well as studying cognitive functions and related diseases. The relationship between the structure of odor molecules and olfactory receptors is complex. Odor molecules with similar structures can react with different olfactory receptors, and odor molecules with different structures can also react with the same olfactory receptors. This complexity of the reaction between odor molecules and olfactory receptors allows the brain to trigger corresponding odors through the activation pattern of olfactory receptor combinations. Traditionally, the study of the interaction between odor molecules and olfactory receptors relies on experimental methods such as bioassays and physicochemical analysis, but this method is often time-consuming, costly and difficult to expand to large-scale sample analysis. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method and device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] In the first aspect, a method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network is provided.

[0007] The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network comprises the following steps:

[0008] Step S1, constructing a first data set, wherein the first data set involves odor molecules that react with olfactory receptors and odor molecules that do not react with olfactory receptors, each odor molecule has multiple physicochemical descriptors, the first data set includes a training set and a test set, the physicochemical descriptors of each odor molecule are used as training data, the training set in the first data set involves two types of odor molecules, the first type of molecules are odor molecules that react with olfactory receptors, and the second type of molecules are odor molecules that do not react with olfactory receptors, the training set in the first data set is used to train a deep learning network model, and the test set in the first data set is used to test the trained deep learning network model, so as to construct a prediction model for odor molecules and olfactory receptor reactions;

[0009] Step S2, based on the odor molecule and olfactory receptor reaction prediction model in step S1, multiple odor molecules that react with olfactory receptors are selected to construct a second data set and a third data set, wherein the odor molecules in the second data set and the third data set respectively have molecular sequence data and molecular structure images, and the molecular sequence data and the molecular structure images are both used as feature representations of the corresponding odor molecules, and the second data set and the third data set also include a training set and a test set, and the molecular sequence data and the molecular structure images of the odor molecules are used as training data, and the training sets in the second data set and the third data set are used to train the network computing model and the deep neural network respectively, and the test sets in the second data set and the third data set are used to test the trained network computing model and the deep neural network respectively, so as to construct the corresponding first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model, and determine the probability distribution of all olfactory receptors corresponding to the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model;

[0010] Step S3, input the set of odor molecules to be tested into the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model respectively, obtain the probability values ​​corresponding to the two sets of probability distributions, add the two probability values ​​and then take the average to obtain the average probability value, set a threshold, and compare the average probability value with the threshold to determine the olfactory receptor that reacts to the odor molecules to be tested.

[0011] In certain embodiments, the physicochemical descriptors include molecular weight, three-dimensional structural characteristics of the molecule, molecular polarity, and lipid solubility parameters;

[0012] Each odor molecule m has a corresponding set of physical and chemical descriptors

[0013]

[0014] Where D is a set of odor molecules and physicochemical descriptors, which includes odor molecules and their corresponding physicochemical descriptors, M is a set of odor molecules that can stimulate olfactory receptors and those that cannot stimulate olfactory receptors, d is the number of physicochemical descriptors, and x is mi is the i-th physicochemical descriptor of odor molecule m.

[0015] In some embodiments, the deep learning network model in step S1 is F(·), W is the network parameter, L(·) is the loss function, and the specific formula is as follows:

[0016]

[0017] in, is the network's predicted output for odor molecule m, |M| is the number of odor molecules contained in the odor molecule set M, is the actual label of the odor molecule, and l(·) is the loss function for a single sample.

[0018] In some embodiments, using the test set in the first data set to test the trained deep learning network model includes:

[0019] The odor molecule m′ to be tested is predicted using the trained deep learning network model F(·):

[0020]

[0021] in, is the predicted output of the odor molecule m′, which can be distinguished by setting an appropriate threshold Get whether the odor molecule m′ can react with the olfactory receptor, is the physicochemical descriptor of the odor molecule m′;

[0022] The deep learning network model includes multiple layers of neurons and can capture the complex relationship between molecular features and olfactory receptor reactivity through nonlinear transformation.

[0023] In some embodiments, constructing a corresponding first odor molecule and olfactory receptor reaction prediction model, and determining the probability distribution of all olfactory receptors based on the first odor molecule and olfactory receptor reaction prediction model includes:

[0024] Step S211, select an odor molecule sample that can react with the olfactory receptor, use the molecular sequence data as its unique feature representation, and identify the olfactory receptor that the odor molecule can react with through the molecular sequence data, as shown in formula (4), I is a data set containing odor molecules and their sequence data, and each odor molecule m has its unique molecular sequence data where [x m1 ,x m2,...x md ] is the molecular physicochemical descriptor, [x md+1 ,x md+2 ,...x mg ] is the molecular fingerprint, and E is the set of odor molecules that can react with the olfactory receptors:

[0025]

[0026] In step S212, the molecular sequence data is used as the feature representation of the odor molecule, and the network computing model learns the complex mapping relationship between the odor molecule features and their corresponding olfactory receptors, so as to accurately identify the olfactory receptors that respond to the odor molecule; as shown in formula (5), Z(·) is the network computing model, and L(·) is the loss function:

[0027]

[0028] Among them, θ is the model parameter, is the molecular sequence data, is the network's predicted output for the odor molecule m, is the actual label of the odor molecule, Indicates whether the odor molecule m can react with the i-th olfactory receptor, N is the set of all olfactory receptors, and |N| is the number of olfactory receptors;

[0029] Step S213, for the odor molecule to be tested, the molecular sequence data of the odor molecule is used as input, and the network computing model is used to identify the specific olfactory receptor that reacts with the odor molecule, as shown in formula (6). For the odor molecule to be tested m′, the trained network model Z(·) is used to make a prediction:

[0030]

[0031] in, is the predicted output of the odor molecule m′, a corresponding first odor molecule and olfactory receptor reaction prediction model is constructed, and the probability distribution of all olfactory receptors based on the first odor molecule and olfactory receptor reaction prediction model is determined.

[0032] In certain embodiments, the molecular structure image displays the planar structural information of the molecule, and the planar structural information of the molecule includes the connection mode between atoms, the type of chemical bonds, and the existence of rings.

[0033] In a second aspect, a device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network is provided. The device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network comprises a memory and one or more processors. The memory stores executable code. When the processor executes the executable code, it is used to implement the method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network as described in any one of the first aspects.

[0034] The beneficial effects of the present invention are as follows: the present invention firstly utilizes the interaction data between odor molecules and olfactory receptors, extracts the physical and chemical descriptors of odor molecules as their features, and utilizes a deep learning network model to determine whether unknown odor molecules can stimulate olfactory receptors. Secondly, the molecular sequence data of odor molecules are used to determine the possibility of reaction between odor molecules and various olfactory receptors through a network calculation model. Then, the molecular structure image data of odor molecules is further introduced to calculate the reaction probability between odor molecules and various olfactory receptors. Finally, in order to improve the accuracy and reliability of the recognition result, the final reaction olfactory receptor of the odor molecule is determined by analyzing the calculated results, i.e., two groups of probability distributions. The present invention can accurately determine whether a molecule can react with an olfactory receptor by using a multimodal deep learning network, and further identify the olfactory receptor with which the odor molecule reacts. It can not only process known molecular structures, but also generalize to unknown or complex molecular structures, thereby improving the applicability of the model. The present method utilizes a calculation method to replace experimental operations, reduces the consumption of laboratory resources, and reduces the cost of studying the reaction between odor molecules and olfactory receptors. Combined with advanced data processing technology and machine learning algorithms, the present invention provides an efficient and accurate method for identifying the reaction between odor molecules and olfactory receptors. In addition, the method of the present invention can be widely used in the fields of drug design, fragrance development, environmental monitoring, etc., providing a new technical means for research and application in related fields, and has wide practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 A flowchart of a method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network provided in an embodiment of the present invention.

[0037] Figure 2 A partial physical and chemical descriptor diagram provided for an embodiment of the present invention.

[0038] Figure 3 This is a partial molecular fingerprint provided by an embodiment of the present invention.

[0039] Figure 4 This is a partial molecular structure image provided by an embodiment of the present invention.

[0040] Figure 5 A diagram of olfactory receptor types provided for embodiments of the present invention.

[0041] Figure 6 The network calculation model provided in the embodiment of the present invention is based on the ROC curve obtained by testing the molecular sequence data and the molecular structure image respectively, and the ROC curve diagram obtained by analyzing the calculation results of the two methods.

[0042] Figure 7 The network calculation model provided for the embodiment of the present invention is based on the true probability value obtained by identifying the olfactory receptor based on the molecular sequence data and the molecular structure image, and the true probability value graph obtained by analyzing the calculation results of the two methods.

[0043] Figure 8 A diagram for identifying olfactory receptor categories provided by an embodiment of the present invention.

[0044] Fig. 9 A schematic diagram of the structure of an apparatus for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to better understand the technical solution of the present application, the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0046] First, define the method for evaluating model performance, using accuracy, true positive rate, and AUROC value for evaluation; evaluation needs to be done from two aspects:

[0047] (1) The model determines whether an odor molecule can react with an olfactory receptor; (2) The model identifies the specific olfactory receptor with which the odor molecule reacts.

[0048] The definitions of accuracy, true positive rate, and AUROC value are as follows:

[0049] To evaluate the performance of the model in determining whether an odor molecule can react with an olfactory receptor, the accuracy is defined as follows:

[0050]

[0051] Among them, k is the number of odor molecules that can correctly predict whether they can react with the olfactory receptors, and n is the total number of test molecules.

[0052] The accuracy, true positive rate and AUROC value were used to evaluate the performance of the model in identifying the specific olfactory receptors that the odor molecules react to.

[0053] 1) The accuracy is defined as follows:

[0054]

[0055] Where n represents the number of tested odor molecules, |N represents the number of olfactory receptor categories, and c i It indicates the prediction of the i-th odor molecule to identify whether the olfactory receptor can react with it, among which the number of correct olfactory receptors is identified, that is, the number of olfactory receptors that can react with the odor molecule and the number of olfactory receptors that cannot react with the odor molecule are correctly identified. The higher the value, the better the overall recognition performance of the model.

[0056] 2) The true rate is defined as follows:

[0057]

[0058] Among them, TP represents the number of true positive examples, that is, the number of responsive olfactory receptors correctly identified by the model, and FN represents the number of false negative examples, that is, the number of responsive olfactory receptors incorrectly identified by the model as non-responsive olfactory receptors. The TPR value is between 0 and 1. The higher the value, the better the model's ability to identify responsive olfactory receptors and the better the performance of the model.

[0059] 3) The AUROC value is defined as follows:

[0060] AUROC (Area Under the Receiver Operating Characteristic Curve) is a statistic that measures the performance of a binary classification model. It represents the area under the ROC curve and its value is between 0 and 1. The higher the value, the better the classification performance of the model. The ROC (Receiver Operating Characteristic curve) curve is a curve representation of the performance of the model under different threshold settings by plotting a graph of the true positive rate against the false positive rate.

[0061] AUROC can be divided into two types: one is the micro-average area under the curve Micro-AUROC, and the other is the macro-average area under the curve Macro-AUROC. Micro-AUROC first calculates the total number of true positives, false positives, and false negatives of all categories, draws the ROC curve based on these totals, and calculates the AUROC value. Micro-AUROC focuses on overall performance and is suitable for processing data sets with imbalanced categories; Macro-AUROC calculates the AUROC value for the ROC curve of each category separately, and then takes the average of these AUROC values. It gives the same weight to all categories, pays attention to the performance of each category, and is suitable for balanced categories. Finally, the larger the values ​​of the micro-average area under the curve Micro-AUROC and the macro-average area under the curve Macro-AUROC, the better the performance of the model.

[0062] We used specific odor molecules and olfactory receptor examples and selected specific deep learning neural networks for actual testing to verify the performance and feasibility of this solution. The specific experimental results are as follows.

[0063] In some embodiments, Figure 1 — Figure 8 As shown, a method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network is provided. The method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network comprises the following steps:

[0064] Step S1, constructing a first data set, wherein the first data set involves odor molecules that react with olfactory receptors and odor molecules that do not react with olfactory receptors, and each odor molecule has multiple physical and chemical descriptors.

[0065] In some embodiments, the physicochemical descriptors are a series of quantitative values ​​used to characterize the physical, chemical, geometric and electronic properties of molecules, including molecular weight, three-dimensional structural characteristics of molecules, molecular polarity, lipid solubility parameters, etc. Thousands of physicochemical descriptor values ​​are usually used to uniquely characterize molecules.

[0066] Each molecule m has a corresponding set of physicochemical descriptors

[0067]

[0068] Where D is a set of odor molecules and physicochemical descriptors, which includes odor molecules and their corresponding physicochemical descriptors, M is a set of odor molecules that can stimulate olfactory receptors and those that cannot stimulate olfactory receptors, d is the number of physicochemical descriptors, and x is mi is the i-th physicochemical descriptor of odor molecule m.

[0069] Figure 2The following are some of the physicochemical descriptors used in this example. The physicochemical descriptors used in this example are obtained by the alvaDesc tool based on the SMILES (Simplified Molecular-Input Line-Entry System) format string of the molecule.

[0070] The first dataset includes a training set and a test set, using the physicochemical descriptors of each odor molecule as training data;

[0071] In some embodiments, 4550 odor molecules are used as training data, including 2000 odor molecules that can react with olfactory receptors and 1550 odor molecules that cannot react with olfactory receptors. Each odor molecule is characterized by 1247 physical and chemical descriptors. The label of each odor molecule is 1 or 0, 1 indicates that the odor molecule can react with the olfactory receptor, and 0 indicates that the odor molecule cannot react with the olfactory receptor.

[0072] The training set in the first data set involves two types of odor molecules. The first type of odor molecules are odor molecules that react with olfactory receptors, and the second type of odor molecules are odor molecules that do not react with olfactory receptors. The training set in the first data set is used to train the deep learning network model to learn a pattern that can distinguish whether an odor molecule can react with an olfactory receptor. During the training process, the deep learning network model simultaneously learns the characteristics of odor molecules that can react with olfactory receptors and those that cannot react with olfactory receptors, thereby enhancing the generalization ability of the model.

[0073] In some embodiments, a deep learning network model such as a ViT (Vision Transformer) network is specifically used for model training, with a ratio of training set to validation set of 8:2, an Adam optimizer, a learning rate of 0.001, a batch size of 64, 2732 iterations, a training time of 10 hours, and a test accuracy of 93% on the validation set.

[0074] The ViT network is an architecture that uses a Transformer model for computer vision tasks. It converts images into serialized data that can be processed by the Transformer by dividing the image into a series of small blocks and linearly mapping these small blocks to a vector sequence of fixed dimensions. These vectors are then processed by an encoder and passed through a multi-layer perceptron head for classification and other downstream tasks. For example, by converting a sequence of digits into a format suitable for processing by the ViT network, each digit is mapped to a high-dimensional space, and the sequence of digits is converted into a series of embedded vectors to achieve data feature extraction and training recognition tasks.

[0075] In some embodiments, the deep learning network model in step S1 is F(·), W is the network parameter, L(·) is the loss function, and the specific formula is as follows:

[0076]

[0077] in, is the network's predicted output for odor molecule m, |M| is the number of odor molecules contained in the odor molecule set M, is the actual label of the odor molecule (1 or 0, i.e., it can react with the olfactory receptor or not), and l(·) is the loss function of a single sample, such as mean square error loss.

[0078] The trained deep learning network model is tested using the test set in the first data set to construct a prediction model for the reaction of odor molecules and olfactory receptors. Based on the physical and chemical descriptor data of the odor molecules to be tested, the reasoning ability of the deep learning network model after training is used to determine whether the odor molecules have the possibility of reacting with the olfactory receptors.

[0079] As shown in formula (3), for the odor molecule m′ to be tested, the trained deep learning network model F(·) is used to make predictions:

[0080]

[0081] in, is the predicted output of the odor molecule m′, which can be distinguished by setting an appropriate threshold Get whether the odor molecule m′ can react with the olfactory receptor, It is the physicochemical descriptor of the odor molecule m′.

[0082] The deep learning network model includes multiple layers of neurons, which can capture the complex relationship between the characteristics of odor molecules and the reactivity of olfactory receptors through nonlinear transformation. In addition, the deep learning network model is optimized through appropriate activation functions and loss functions to ensure the accuracy and reliability of the prediction results. The deep learning network model F(·) extracts the characteristics of odor molecules and establishes a prediction model by learning a large amount of data D on the interaction between odor molecules and olfactory receptors. The training goal of the model is to minimize the loss function L(·) and improve the accuracy of judging the reactivity of odor molecules.

[0083] Model testing: The test set uses 1,362 odor molecules that were not used in the training, of which 700 can react with olfactory receptors and 662 cannot. The accuracy on the test set is 87%, which means that 87% of the odor molecules can be correctly predicted to react with olfactory receptors.

[0084] Step S2, based on the odor molecule and olfactory receptor reaction prediction model in step S1, multiple odor molecules that react with olfactory receptors are selected to construct a second data set and a third data set, wherein the odor molecules in the second data set and the third data set respectively have molecular sequence data and molecular structure images, and the molecular sequence data and the molecular structure images are both used as feature representations of the corresponding odor molecules. The second data set and the third data set also include a training set and a test set, and the molecular sequence data and the molecular structure images of the odor molecules are used as training data. The training sets in the second data set and the third data set are used to train the network computing model and the deep neural network respectively, and the test sets in the second data set and the third data set are used to test the trained network computing model and the deep neural network respectively, so as to construct the corresponding first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model, and determine the probability distribution of all olfactory receptors corresponding to the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model;

[0085] In some embodiments, using molecular sequence data to determine the likelihood of a reaction between an odor molecule and each olfactory receptor through a network computational model includes the following steps:

[0086] Step S211, selecting an odor molecule sample that can react with an olfactory receptor;

[0087] In some embodiments, for each odor molecule, 1247 physicochemical descriptors and 166 molecular fingerprint data are used as molecular sequence data to further supplement the characteristic information of the odor molecule, and the molecular sequence data is extracted from the odor molecule using the PaDEL tool. Figure 3 This is the partial molecular sequence data used in this example.

[0088] The molecular sequence data is used as its unique feature representation, and the olfactory receptors that the odor molecule can react with are identified through the molecular sequence data, as shown in formula (4), where I is a data set containing odor molecules and their sequence data, and each odor molecule m has its unique molecular sequence data where [x m1 ,x m2 ,...x md ] is the molecular physicochemical descriptor, [x md+1 ,x md+2 ,...x mg ] is the molecular fingerprint, and E is the set of odor molecules that can react with the olfactory receptors:

[0089]

[0090] In step S212, the molecular sequence data is used as the feature representation of the odor molecule, and the network computing model learns the complex mapping relationship between the molecular features and their corresponding olfactory receptors, so as to accurately identify the olfactory receptors that react to the odor molecule; as shown in formula (5), Z(·) is the network computing model, and L(·) is the loss function:

[0091]

[0092] Among them, θ is the model parameter, is the molecular sequence data, is the network's predicted output for the odor molecule m, is the actual label of the odor molecule, Indicates whether the odor molecule m can react with the i-th olfactory receptor, N is the set of all olfactory receptors, and |N| is the number of olfactory receptors;

[0093] In some embodiments, the sequence data of each odor molecule is used as a feature representation of the molecule and as training data for the network computing model. The example specifically uses 3240 odor molecules, all of which can react with olfactory receptors, corresponding to a total of 52 olfactory receptors, and each odor molecule corresponds to one or more olfactory receptors. The extracted molecular sequence data is used as input features and trained using a DNN network. The ratio of the training set to the validation set is 8:2, using the Adam optimizer, the learning rate is 0.001, the batch size is 64, the number of iterations is 3742, the training time is 6 hours, and the test accuracy on the validation set is 91%, indicating that the model can effectively learn the association between molecular sequence features and reactive olfactory receptors.

[0094] Model testing: We used odor molecules that were not involved in the training to test the model and verify the model's ability to distinguish molecular response olfactory receptors. We used 1,356 odor molecules that were not involved in the training to test, with an accuracy of 86%, verifying the model's generalization ability on unknown data. We also used the micro-average area under the curve and the macro-average area under the curve to evaluate the model performance. Figure 6 (a) is the ROC curve of the example network for the test odor molecules, where the micro-average area under the curve is 0.873 and the macro-average area under the curve is 0.842, indicating that the trained network has a good ability to distinguish the olfactory receptors that respond to odor molecules. Figure 7 (a) is the true positive rate value obtained by the network in this example to identify 52 olfactory receptors, and it has good prediction results for most olfactory receptors.

[0095] Complete the construction of the corresponding first odor molecule and olfactory receptor response prediction model.

[0096] Step S213, for the odor molecule to be tested, the molecular sequence data of the odor molecule is used as input, and the network computing model is used to identify the specific olfactory receptor that reacts with the odor molecule, as shown in formula (6). For the odor molecule to be tested m′, the trained network computing model Z(·) is used to make a prediction:

[0097]

[0098] in, is the predicted output of odor molecule m′;

[0099] Determine the probability distribution of all olfactory receptors based on the prediction model of the reaction of the first odor molecule with the olfactory receptor.

[0100] In some embodiments, using molecular structure image data, determining the possibility of a reaction between an odor molecule and each olfactory receptor through a deep neural network includes the following steps:

[0101] Step S221, select odor molecules that can react with olfactory receptors as a data set, and use molecular structure images for feature representation of odor molecules. Molecular structure images can capture the planar structural information of molecules, such as the connection mode and chemical bond type between molecules, as well as the atomic type, chemical bond type and ring characteristics, etc., and use them as the features of each molecule and as training data for the deep neural network. In this example, the RDKit tool is used to obtain the molecular structure image from the SMILES format string of the molecule;

[0102] Step S222, the structural image of each odor molecule is used as the training data of the deep neural network, and 3240 odor molecules are used as training data, corresponding to 52 olfactory receptors in total, and each odor molecule corresponds to one or more olfactory receptors. The example uses the ConvNeXt network for model training, the ratio of the training set to the validation set is 8:2, the Adam optimizer is used, the learning rate is 0.001, the batch size is 64, the number of iterations is 3651, and the training time is 7.5 hours. The test accuracy on the validation set is 93%, indicating that the model has a good ability to distinguish the specific response olfactory receptors of the odor molecules.

[0103] The ConvNext network combines the advantages of convolutional neural networks and Transformer to improve the performance of image classification tasks by adjusting the calculation ratio of different stages to match the pattern in Transformer, and using depthwise separable convolution technology and large convolution kernel size to reduce the number of parameters and computation while maintaining efficiency.

[0104] Model testing: 1,356 odor molecules that were not used in training were used for testing, with a test accuracy of 88%. Figure 6(b) is the ROC curve of the example network for the test molecule, where the micro-average area under the curve is 0.902 and the macro-average area under the curve is 0.863, indicating that the network has good recognition and discrimination capabilities for the olfactory receptor reactivity of odor molecules. Figure 7 (b) The true positive rate value obtained by the network in this example for identifying 52 olfactory receptors, with good prediction results for most olfactory receptors.

[0105] Complete the construction of the corresponding second odor molecule and olfactory receptor reaction prediction model.

[0106] Step S223: For the odor molecule to be tested, the deep neural network analyzes the molecular structure image features to obtain its corresponding olfactory receptor. For the odor molecule to be tested m′, the trained deep neural network G(·) is used for prediction and recognition:

[0107]

[0108] Among them, ω is the model parameter, s m′ is the structural image of the odor molecule m′, is the predicted output result of odor molecule m′.

[0109] Determine the probability distribution of all olfactory receptors based on the prediction model of the reaction of the second odor molecule with the olfactory receptor.

[0110] Step S3, input the set of molecules to be tested into the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model respectively, obtain the probability values ​​corresponding to the two sets of probability distributions, add the two probability values ​​and then take the average to obtain the average probability value, set a threshold, and compare the average probability value with the threshold to determine the olfactory receptor that reacts to the odor molecule to be tested.

[0111] In some embodiments, the specific steps of determining the olfactory receptor that responds to the odor molecule to be tested are as follows:

[0112] Step S31: for the odor molecule m′ to be tested, the network calculation model Z(·) is used to predict the probability distribution of all olfactory receptors based on the molecular sequence data. Using a deep neural network G(·), we predict the probability distribution of all olfactory receptors based on the molecular structure images.

[0113] Step S32, for the probability distributions of the two groups of olfactory receptors, the corresponding probability values ​​are added, and then the average is taken to obtain the probability distribution p of all olfactory receptors in another group. m′ , set a suitable threshold T, and select the final set of olfactory receptors R from it. As shown in formula (8):

[0114]

[0115] in, and They represent the probability values ​​corresponding to the i-th olfactory receptor predicted by Z(·) and G(·), respectively.

[0116] In some embodiments, based on the molecular sequence data, a DNN network is used to identify a set of probabilities that the odor molecule reacts with all olfactory receptors; then, based on the molecular structure image, a ConvNeXt network is used to identify another set of probabilities that the odor molecule reacts with all olfactory receptors. For the two sets of reaction probabilities, the corresponding probability values ​​are added and averaged, and a threshold is set to 0.5 to determine the olfactory receptor that reacts to the odor molecule.

[0117] Using 1,356 odor molecules for testing, the accuracy rate increased from 86% and 88% to 93%. Figure 6 (c) is the ROC curve of this method for olfactory receptors, where the micro-average area under the curve is 0.962 and the macro-average area under the curve is 0.924. Figure 7 (c) The true value of each olfactory receptor tested by this method is significantly improved compared with the olfactory receptor that only uses a single method to judge the response of odor molecules.

[0118] The present invention firstly utilizes the interaction data between odor molecules and olfactory receptors, extracts the physical and chemical descriptors of odor molecules as their features, and utilizes a deep learning network model to determine whether unknown odor molecules can stimulate olfactory receptors. Secondly, the molecular sequence data of odor molecules are used to determine the possibility of reaction between odor molecules and various olfactory receptors through a network calculation model. Then, the molecular structure image data of odor molecules is further introduced to calculate the reaction probability between odor molecules and various olfactory receptors. Finally, in order to improve the accuracy and reliability of the recognition result, the final reaction olfactory receptor of the odor molecule is determined by analyzing the calculated results, that is, two groups of probability distributions. The present invention can accurately determine whether odor molecules can react with olfactory receptors by using a multimodal deep learning network, and further identify the olfactory receptors with which the odor molecules react. It can not only process known molecular structures, but also generalize to unknown or complex molecular structures, thereby improving the applicability of the model. The present method utilizes a calculation method to replace experimental operations, reduces the consumption of laboratory resources, and reduces the cost of studying the reaction between odor molecules and olfactory receptors. Combined with advanced data processing technology and machine learning algorithms, the present invention provides an efficient and accurate method for identifying the reaction between odor molecules and olfactory receptors. In addition, the method of the present invention can be widely used in the fields of drug design, fragrance development, environmental monitoring, etc., providing a new technical means for research and application in related fields, and has wide practical value.

[0119] Corresponding to the aforementioned embodiment of the method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network, the present invention also provides an embodiment of a device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network.

[0120] See also Fig. 9 An embodiment of the present invention provides a device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network, comprising a memory and one or more processors. The memory stores executable code. When the processor executes the executable code, it is used to implement the method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network in the above embodiment.

[0121] The embodiment of the device for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network of the present invention can be applied to any device with data processing capabilities, and the device with data processing capabilities can be a device or apparatus such as a computer. The device embodiment can be implemented through software, or through hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capabilities in which it is located reading the corresponding computer program instructions in the non-volatile memory into the internal memory for execution. From a hardware perspective, if Fig. 9 As shown, it is a hardware structure diagram of any device with data processing capability where the odor molecule and olfactory receptor reaction prediction device based on multimodal deep learning network of the present invention is located, except Fig. 9 In addition to the processor, memory, network interface, and non-volatile memory shown, any device with data processing capabilities in which the apparatus in the embodiments is located may also include other hardware, generally based on the actual functions of the device with data processing capabilities, which will not be described in detail.

[0122] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0123] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, in which the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0124] An embodiment of the present invention also provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network in the above embodiment is implemented.

[0125] The computer-readable storage medium may be an internal storage unit of any device with data processing capability in any of the aforementioned embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be an external storage device of any device with data processing capability, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing capability. The computer-readable storage medium is used to store computer programs and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0126] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0127] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0128] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] The above are only preferred embodiments of one or more embodiments of this specification, and are not intended to limit one or more embodiments of this specification. Any modifications made within the spirit and principles of one or more embodiments of this specification are not intended to limit this specification.

Claims

1. A method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network, characterized in that: The following steps are involved: Step S1, constructing a first data set, wherein the first data set involves odor molecules that react with olfactory receptors and odor molecules that do not react with olfactory receptors, each odor molecule has multiple physicochemical descriptors, the first data set includes a training set and a test set, the physicochemical descriptors of each odor molecule are used as training data, the training set in the first data set involves two types of odor molecules, the first type of molecules are odor molecules that react with olfactory receptors, and the second type of molecules are odor molecules that do not react with olfactory receptors, the training set in the first data set is used to train a deep learning network model, and the test set in the first data set is used to test the trained deep learning network model, so as to construct a prediction model for odor molecules and olfactory receptor reactions; Step S2, based on the odor molecule and olfactory receptor reaction prediction model in step S1, multiple odor molecules that react with olfactory receptors are selected to construct a second data set and a third data set, wherein the odor molecules in the second data set and the third data set respectively have molecular sequence data and molecular structure images, and the molecular sequence data and the molecular structure images are both used as feature representations of the corresponding odor molecules, and the second data set and the third data set also include a training set and a test set, and the molecular sequence data and the molecular structure images of the odor molecules are used as training data, and the training sets in the second data set and the third data set are used to train the network computing model and the deep neural network respectively, and the test sets in the second data set and the third data set are used to test the trained network computing model and the deep neural network respectively, so as to construct the corresponding first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model, and determine the probability distribution of all olfactory receptors corresponding to the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model; Step S3, input the set of odor molecules to be tested into the first odor molecule and olfactory receptor reaction prediction model and the second odor molecule and olfactory receptor reaction prediction model respectively, obtain the probability values ​​corresponding to the two sets of probability distributions, add the two probability values ​​and then take the average to obtain the average probability value, set a threshold, and compare the average probability value with the threshold to determine the olfactory receptor that reacts to the odor molecules to be tested.

2. The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network according to claim 1, characterized in that: The physicochemical descriptors include molecular weight, three-dimensional structural characteristics of the molecule, molecular polarity and lipid solubility parameters; Each odor molecule m has a corresponding set of physical and chemical descriptors Where D is a set of odor molecules and physicochemical descriptors, which includes odor molecules and their corresponding physicochemical descriptors, M is a set of odor molecules that can stimulate olfactory receptors and those that cannot stimulate olfactory receptors, d is the number of physicochemical descriptors, and x is m i is the i-th physicochemical descriptor of odor molecule m.

3. The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network according to claim 2, characterized in that: In step S1, the deep learning network model is F(·), W is the network parameter, L(·) is the loss function, and the specific formula is as follows: in, is the network's predicted output for odor molecule m, |M| is the number of odor molecules contained in the odor molecule set M, is the actual label of the odor molecule, and l(·) is the loss function for a single sample.

4. The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network according to claim 3, characterized in that: Using the test set in the first data set to test the trained deep learning network model includes: The odor molecule m′ to be tested is predicted using the trained deep learning network model F(·): in, is the predicted output of the odor molecule m′, which can be distinguished by setting an appropriate threshold Get whether the odor molecule m′ can react with the olfactory receptor, is the physicochemical descriptor of the odor molecule m′; The deep learning network model includes multiple layers of neurons and can capture the complex relationship between odor molecule characteristics and olfactory receptor reactivity through nonlinear transformation.

5. The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network according to claim 2, characterized in that: Constructing a corresponding prediction model of the reaction between the first odor molecule and the olfactory receptor, and determining the probability distribution of all olfactory receptors based on the prediction model of the reaction between the first odor molecule and the olfactory receptor includes: Step S211, select an odor molecule sample that can react with the olfactory receptor, use the odor molecule sequence data as its unique feature representation, and identify the olfactory receptor with which the odor molecule can react through the odor molecule sequence data, as shown in formula (4), I is a data set containing odor molecules and their sequence data, and each odor molecule m has its unique molecular sequence data where [x m1 ,x m2 ,...x md ] is the molecular physicochemical descriptor, [x md+1 ,x md+2 ,...x mg ] is the molecular fingerprint, and E is the set of odor molecules that can react with the olfactory receptors: In step S212, the molecular sequence data is used as the feature representation of the odor molecule, and the network computing model learns the complex mapping relationship between the odor molecule features and their corresponding olfactory receptors, so as to accurately identify the olfactory receptors that respond to the odor molecule; as shown in formula (5), Z(·) is the network computing model, and L(·) is the loss function: Among them, θ is the model parameter, is the molecular sequence data, is the network’s predicted output for the odor molecule m, is the actual label of the odor molecule, Indicates whether the odor molecule m can react with the i-th olfactory receptor, N is the set of all olfactory receptors, and |N| is the number of olfactory receptors; Step S213, for the odor molecule to be tested, the molecular sequence data of the odor molecule is used as input, and the network computing model is used to identify the specific olfactory receptor that reacts with the odor molecule, as shown in formula (6). For the odor molecule to be tested m′, the trained network model Z(·) is used to make a prediction: in, is the predicted output of the odor molecule m′, a corresponding first odor molecule and olfactory receptor reaction prediction model is constructed, and the probability distribution of all olfactory receptors based on the first odor molecule and olfactory receptor reaction prediction model is determined.

6. The method for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network according to claim 1, characterized in that: The molecular structure image shows the planar structural information of the molecule, and the planar structural information of the molecule includes the connection mode between atoms, the type of chemical bonds and the existence of rings.

7. A device for predicting the reaction between odor molecules and olfactory receptors based on a multimodal deep learning network, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the processor executes the executable code, it is used to implement the method for predicting the reaction of odor molecules and olfactory receptors based on a multimodal deep learning network as described in any one of claims 1 to 6.

Images