Mass spectrum search matching method and device based on deep learning and storage medium

Abstract

The application discloses a mass spectrum search matching method and device based on deep learning and a storage medium. The application divides a data set into a training set, a verification set and a test set. Data enhancement is performed on the training set, the verification set and the test set. Molecular fingerprints of the training set are extracted, and similarity scores of the molecular fingerprints are calculated. The similarity scores are used as real similarity scores of two mass spectra. The two mass spectra are converted into mass spectrum vectors, and the two mass spectrum vectors are respectively input into the same feedforward neural network. Hidden features of the mass spectra are extracted, and high-order vector representations of the mass spectra are obtained. The two mass spectrum vectors are interacted through a similarity calculation layer, and similarity prediction scores of the two mass spectra are obtained. The application can improve the success rate of mass spectrum matching and the accuracy of component matching based on the ability of the enhanced method in the result of a mass spectrum database search engine.

Description

Technical Field

[0001] This application relates to the field of mass spectrum retrieval and analysis, and in particular to mass spectrum search and matching methods, devices and storage media based on deep learning. Background Technology

[0002] Manual analysis of mass spectrometry data remains a significant challenge in Vocs mass spectrometry research, particularly as manually validating the results of mass spectrometry database search algorithms can be extremely time-consuming. While mass spectrometry similarity metrics are often used as proxies for structural similarity, this approach is limited by the correlation between the two metrics.

[0003] Therefore, the aforementioned technical problems in the relevant technologies urgently need to be solved. Summary of the Invention

[0004] This application aims to solve one of the technical problems in related technologies. To this end, embodiments of this application provide a deep learning-based mass spectrum search and matching method, apparatus, and storage medium, which can improve the accuracy of mass spectrum component matching.

[0005] According to one aspect of the embodiments of this application, a mass spectrometry search and matching method based on deep learning is provided, the method comprising:

[0006] The dataset is divided into training set, validation set, and test set;

[0007] Data augmentation is performed on the training set, validation set, and test set;

[0008] Extract the molecular fingerprint from the training set, calculate the similarity score of the molecular fingerprint, and use the similarity score as the true similarity score between the two mass spectra.

[0009] The two mass spectra are converted into mass spectrum vectors, and the two mass spectrum vectors are respectively input into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra.

[0010] The two mass spectrum vectors are interacted through a similarity calculation layer to obtain the similarity prediction score between the two mass spectra.

[0011] In one embodiment, before dividing the dataset into training, validation, and test sets, the method further includes data preparation, including:

[0012] The dataset used was annotated MS / MS spectra collected in GNPS;

[0013] The metadata and mass spectrometry peak data of the dataset were processed using the matchms toolkit.

[0014] In one embodiment, the metadata of the dataset is processed, including:

[0015] Extract the summation information from the metadata;

[0016] The metadata was moved to a consistent field, and the pubchempy toolkit was used to automatically search for and complete the spectra with missing SMILES annotations on PubChem based on the additive information.

[0017] In one embodiment, the mass spectrometry peak data of the dataset is processed, including:

[0018] The units m / z of the mass spectrometry peak data are converted into intensity indices;

[0019] Remove peaks with an intensity lower than a preset intensity, and keep the number of the highest peaks within a preset limit;

[0020] The number of peaks with atomic mass between the first and second preset values ​​is greater than or equal to the third preset value.

[0021] In one embodiment, data augmentation is performed on the training set, validation set, and test set, including:

[0022] Low-intensity peak removal: Randomly select 0-20% of peaks with an intensity below 0.4 and set them to 0;

[0023] Peak jitter: The peak of each non-zero intensity peak varies randomly between 0 and ±40%;

[0024] Peak addition: Randomly select 0 to 10 peaks with an intensity value of 0, and randomly set the value to 0 to 0.01.

[0025] In one embodiment, calculating the similarity score of the molecular fingerprint includes:

[0026] The similarity score of molecular fingerprints is calculated using the Tanimoto coefficient.

[0027] In one embodiment, after obtaining the similarity prediction scores of the two mass spectra, the method further includes:

[0028] The similarity prediction score is used to determine whether the two mass spectra match.

[0029] According to one aspect of an embodiment of this application, a mass spectrometry search and matching device based on deep learning is provided, the device comprising:

[0030] The first module is used to divide the dataset into training set, validation set, and test set;

[0031] The second module is used to perform data augmentation on the training set, validation set, and test set.

[0032] The third module is used to extract the molecular fingerprint of the training set, calculate the similarity score of the molecular fingerprint, and use the similarity score as the true similarity score of the two mass spectra.

[0033] The fourth module is used to convert the two mass spectra into mass spectra vectors, and input the two mass spectra vectors into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra.

[0034] The fifth module is used to interact two mass spectrum vectors through a similarity calculation layer to obtain the similarity prediction score between the two mass spectra.

[0035] According to one aspect of an embodiment of this application, a mass spectrometry search and matching device based on deep learning is provided, the device comprising:

[0036] At least one processor;

[0037] At least one memory for storing at least one program;

[0038] When at least one of the programs is executed by at least one of the processors, the deep learning-based mass spectrometry search and matching method described in the preceding embodiments is implemented.

[0039] According to one aspect of the embodiments of this application, a storage medium is provided, the storage medium storing a processor-executable program, which, when executed by a processor, implements the deep learning-based mass spectrometry search and matching method as described in the preceding embodiments.

[0040] The beneficial effects of the deep learning-based mass spectrometry search and matching method, apparatus, and storage medium provided in this application are as follows: This application divides the dataset into a training set, a validation set, and a test set; performs data augmentation on the training set, validation set, and test set; extracts the molecular fingerprint of the training set, calculates the similarity score of the molecular fingerprint, and uses the similarity score as the true similarity score between the two mass spectra; converts the two mass spectra into mass spectrometry vectors, inputs the two mass spectrometry vectors into the same feedforward neural network, extracts the latent features of the mass spectra, and obtains the high-order vector representation of the mass spectra; and interacts the two mass spectrometry vectors through a similarity calculation layer to obtain the predicted similarity score between the two mass spectra. This application, based on the ability to augment the results of a mass spectrometry database search engine, can improve the success rate of mass spectrometry matching and the accuracy of component matching.

[0041] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart of a deep learning-based mass spectrum search and matching method provided for embodiments of this application;

[0044] Figure 2 A diagram of the intelligent mass spectrometry search model provided in the embodiments of this application;

[0045] Figure 3 A schematic diagram of a deep learning-based mass spectrum search and matching device provided in an embodiment of this application;

[0046] Figure 4 This is a schematic diagram of another deep learning-based mass spectrum search and matching device provided in an embodiment of this application. Detailed Implementation

[0047] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0048] The terms "first," "second," "third," and "fourth," etc., used in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0049] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0050] Manual analysis of mass spectrometry data remains a significant challenge in Vocs mass spectrometry research, particularly as manually validating the results of mass spectrometry database search algorithms can be extremely time-consuming. While mass spectrometry similarity metrics are often used as proxies for structural similarity, this approach is limited by the correlation between the two metrics.

[0051] To address the aforementioned issues, this application proposes a deep learning-based mass spectrometry search and matching method, apparatus, and storage medium. This application utilizes deep learning to improve the accuracy of these methods and adapts the model to new data features. Specifically, compared to thresholding and other machine learning methods, this application enhances the method's ability to perform component matching based on the results of a Vocs mass spectrometry database search engine, thereby increasing the success rate of Vocs mass spectrometry matching.

[0052] Specifically, this application proposes a deep learning-based deep network model: a novel deep neural network model that predicts the structural similarity between two chemical structures based on Vocs ion fragment mass spectrometry. Using a clean dataset of >100,000 mass spectra of approximately 15,000 unique known compounds, the DeepVocsMatch model is trained to predict the structural similarity scores of mass spectra pairs with high accuracy. On 1170 mass spectra of 117 Vocs organic compounds, DeepVocsMatch is able to identify highly reliable Vocs structure matches. Furthermore, the uncertainty estimate of the prediction can be used to select a subset of predictions with a root mean square error of approximately 0.1. The model outperforms classical mass spectrometry similarity metrics in retrieving chemically related compound pairs from the Vocs mass spectrometry dataset, demonstrating its capability in Vocs mass spectrometry library matching. The specific processing flow is as follows:

[0053] Figure 1 A flowchart of a deep learning-based mass spectrometry search and matching method provided in the embodiments of this application is shown below. Figure 1 As shown, the deep learning-based mass spectrometry search and matching method proposed in this application includes:

[0054] S101. Divide the dataset into training set, validation set, and test set.

[0055] S102. Perform data augmentation on the training set, validation set, and test set.

[0056] S103. Extract the molecular fingerprint of the training set, calculate the similarity score of the molecular fingerprint, and use the similarity score as the true similarity score of the two mass spectra.

[0057] S104. Convert the two mass spectra into mass spectrum vectors, and input the two mass spectrum vectors into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra.

[0058] S105. The two mass spectrum vectors are interacted through a similarity calculation layer to obtain the similarity prediction scores of the two mass spectra.

[0059] In this embodiment, before dividing the dataset into training, validation, and test sets, step S101 further includes data preparation, including: using annotated MS / MS spectra collected in GNPS as the dataset; and processing the metadata and mass spectrometry peak data of the dataset using the matchms toolkit. Specifically, processing the metadata of the dataset includes: extracting additive information from the metadata; moving the metadata to a consistent field; and using the pubchempy toolkit to automatically search for and complete spectra with missing SMILES annotations on PubChem based on the additive information.

[0060] Accordingly, in this embodiment, the mass spectrometry peak data of the dataset is processed, including: converting the unit m / z of the mass spectrometry peak data into an intensity index; removing peaks with intensities less than a preset intensity, and retaining the number of the highest peaks within a preset number; ensuring that the number of peaks with atomic masses between a first preset value and a second preset value is greater than or equal to a third preset value. For example, for the mass spectrometry peak data, m / z is first converted into an intensity index, then peaks with intensities less than 0.1% are removed, and the number of the highest peaks is retained to within 1000. In addition, the number of peaks with atomic masses between 10 and 1000 Da needs to be greater than or equal to 5. This ultimately produces 109,734 spectra, containing 15,062 unique molecular fingerprints.

[0061] Specifically, to avoid overfitting, data augmentation is required on the training set. Step S102 involves data augmentation of the training set, validation set, and test set, including: low-intensity peak removal: randomly selecting 0-20% of peaks with an intensity below 0.4 and setting them to 0; peak jittering: randomly varying the intensity of each non-zero peak by 0% to ±40%; peak addition: randomly selecting 0 to 10 peaks with an intensity value of 0 and randomly setting their values ​​to 0 to 0.01.

[0062] Step S103, calculating the similarity score of the molecular fingerprint, includes: calculating the similarity score of the molecular fingerprint using the Tanimoto coefficient. After obtaining the similarity prediction scores of the two mass spectra, the method further includes: determining whether the two mass spectra match based on the similarity prediction scores.

[0063] Specifically, the deep learning model training process is as follows: Molecular fingerprints are extracted, and the structural similarity of these fingerprints is calculated using the Tanimoto coefficient. The resulting similarity score is used as the true similarity score between the two mass spectra. The model input consists of two mass spectra files. After preprocessing, the two mass spectra are first converted into vector representations. These vectors are then fed into the same feedforward neural network to extract the latent features of the mass spectra, obtaining higher-order vector representations. Finally, the two mass spectra vector representations are interacted through a similarity calculation layer (fully connected layer, click, or similarity function) to obtain the predicted similarity score between the two mass spectra. The algorithm uses mean squared error as the loss function to calculate the loss score between the true and predicted similarity values. The Adam algorithm is used as the optimizer, optimizing the model parameters through backpropagation.

[0064] This application proposes a deep learning-based Vocs mass spectrum retrieval technique for predicting structural similarity scores between Vocs ion fragment mass spectra. Results show that this method can infer structural similarity between mass spectra with high overall accuracy without requiring any additional metadata or library data. The accuracy of predictions is significantly improved by using deep learning, particularly by merging the predicted scores of mass spectra belonging to the same Vocs organic compound pair. This deep learning-based Vocs mass spectrum search technique will become a powerful new tool for comparing and analyzing Vocs mass spectra, thus having significant implications for the monitoring and analysis of Vocs organic compounds in the environment.

[0065] like Figure 2 As shown, Figure 2 The mass spectrum intelligent search model diagram provided in this application embodiment uses a cleaned dataset of >100,000 mass spectra of approximately 15,000 unique known compounds to train the DeepVocsMatch model to predict the structural similarity score of mass spectrum pairs with high accuracy. On 1170 mass spectra of 117 Vocs organic compounds, DeepVocsMatch was able to identify highly reliable Vocs structure matches.

[0066] In addition, this application also provides a deep learning-based mass spectrum search and matching device, such as... Figure 3 As shown, the device includes:

[0067] The first module is used to divide the dataset into training set, validation set, and test set;

[0068] The second module is used to perform data augmentation on the training set, validation set, and test set.

[0069] The third module is used to extract the molecular fingerprint of the training set, calculate the similarity score of the molecular fingerprint, and use the similarity score as the true similarity score of the two mass spectra.

[0070] The fourth module is used to convert the two mass spectra into mass spectra vectors, and input the two mass spectra vectors into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra.

[0071] The fifth module is used to interact two mass spectrum vectors through a similarity calculation layer to obtain the similarity prediction score between the two mass spectra.

[0072] In addition, this application also provides a deep learning-based mass spectrum search and matching device, such as... Figure 4 As shown, the device includes:

[0073] At least one processor;

[0074] At least one memory for storing at least one program;

[0075] When at least one of the programs is executed by at least one of the processors, the deep learning-based mass spectrometry search and matching method described in the preceding embodiments is implemented.

[0076] In addition, this application also provides a storage medium storing a processor-executable program, which, when executed by a processor, implements the deep learning-based mass spectrometry search and matching method as described in the preceding embodiments.

[0077] Similarly, the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0078] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this application are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0079] Furthermore, although this application is described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding this application. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional technology for an engineer. Therefore, those skilled in the art can implement the application set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of this application, which is determined by the full scope of the appended claims and their equivalents.

[0080] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0081] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0082] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0083] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0084] In the foregoing description of this specification, the references to terms such as "one embodiment," "another embodiment," or "some embodiments," etc., indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0085] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

[0086] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A mass spectrometry search and matching method based on deep learning, characterized in that, The method includes: The dataset is divided into training set, validation set, and test set; Data augmentation is performed on the training set, validation set, and test set; wherein, the data augmentation methods include low-intensity peak removal, peak jittering, and peak addition; Molecular fingerprints are extracted from the training set, and similarity scores are calculated for the molecular fingerprints. These similarity scores are used as the true similarity scores between two mass spectra. The true similarity scores are used to guide the training of the feedforward neural network so that the feedforward neural network can learn the structural similarity between two chemical structures. The two mass spectra are converted into mass spectrum vectors, and the two mass spectrum vectors are respectively input into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra. The two mass spectrum vectors are interacted through a similarity calculation layer to obtain the similarity prediction score between the two mass spectra; Before dividing the dataset into training, validation, and test sets, the method further includes a step of processing the mass spectrometry peak data of the dataset. This processing of the mass spectrometry peak data includes: The units m / z of the mass spectrometry peak data are converted into intensity indices; Remove peaks with an intensity lower than a preset intensity, and keep the number of the highest peaks within a preset limit; The number of peaks with atomic mass between the first and second preset values ​​is greater than or equal to the third preset value.

2. The deep learning-based mass spectrometry search and matching method according to claim 1, characterized in that, Before dividing the dataset into training, validation, and test sets, the method also includes data preparation, including: The dataset used was annotated MS / MS spectra collected in GNPS; The metadata and mass spectrometry peak data of the dataset were processed using the matchms toolkit.

3. The deep learning-based mass spectrometry search and matching method according to claim 2, characterized in that, Processing the metadata of the dataset includes: Extract the summation information from the metadata; The metadata was moved to a consistent field, and the pubchempy toolkit was used to automatically search for and complete the spectra with missing SMILES annotations on PubChem based on the additive information.

4. The deep learning-based mass spectrometry search and matching method according to claim 1, characterized in that, Data augmentation is performed on the training set, validation set, and test set, including: Low-intensity peak removal: Randomly select 0-20% of peak values ​​with an intensity below 0.4 and set them to 0; Peak jitter: The peak of each non-zero intensity peak is randomly varied between 0 and ±40%; Peak addition: Randomly select 0 to 10 peaks with an intensity value of 0, and randomly set the value to 0 to 0.

01.

5. The deep learning-based mass spectrometry search and matching method according to claim 1, characterized in that, The calculation of the similarity score of the molecular fingerprint includes: The similarity score of molecular fingerprints is calculated using the Tanimoto coefficient.

6. The deep learning-based mass spectrometry search and matching method according to claim 1, characterized in that, After obtaining the similarity prediction scores of the two mass spectra, the method further includes: The similarity prediction score is used to determine whether the two mass spectra match.

7. A mass spectrometry search and matching device based on deep learning, characterized in that, The device includes: The first module is used to divide the dataset into training set, validation set, and test set; The second module is used to perform data augmentation on the training set, validation set, and test set; wherein the data augmentation methods include low-intensity peak removal, peak jittering, and peak addition; The third module is used to extract the molecular fingerprint of the training set, calculate the similarity score of the molecular fingerprint, and use the similarity score as the true similarity score of the two mass spectra; wherein, the true similarity score is used to guide the training of the feedforward neural network so that the feedforward neural network learns the structural similarity between the two chemical structures; The fourth module is used to convert the two mass spectra into mass spectra vectors, and input the two mass spectra vectors into the same feedforward neural network to extract the hidden features of the mass spectra and obtain the high-order vector representation of the mass spectra. The fifth module is used to interact two mass spectrum vectors through a similarity calculation layer to obtain the similarity prediction score between the two mass spectra; Before dividing the dataset into training, validation, and test sets, the device further processes the mass spectrometry peak data of the dataset, including: The units m / z of the mass spectrometry peak data are converted into intensity indices; Remove peaks with an intensity lower than a preset intensity, and keep the number of the highest peaks within a preset limit; The number of peaks with atomic mass between the first and second preset values ​​is greater than or equal to the third preset value.

8. A mass spectrum search and matching device based on deep learning, characterized in that, The device includes: At least one processor; At least one memory for storing at least one program; The deep learning-based mass spectrum search and matching method as described in any one of the above-mentioned programs is implemented when at least one of the above-mentioned programs is executed by at least one of the above-mentioned processors.

9. A storage medium, characterized in that, The storage medium stores a processor-executable program, which, when executed by the processor, implements the deep learning-based mass spectrum search and matching method as described in any one of claims 1-6.

Images