Methods, devices, equipment, and media for analyzing multi-target proteins in obesity based on deep learning.

Abstract

This application relates to the interdisciplinary field of deep learning and protein analysis, specifically to a method, apparatus, device, and medium for analyzing multi-target proteins in obesity based on deep learning. The method includes: acquiring biological sample data and preprocessing the biological sample data to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features; constructing an initial deep learning model for analyzing multi-target proteins in obesity based on a multilayer feedforward neural network; training the initial deep learning model based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity; and determining multiple protein expression features related to obesity based on the deep learning model for analyzing multi-target proteins in obesity.

Description

Technical Field

[0001] This disclosure generally relates to the interdisciplinary field of deep learning and protein analysis, and specifically to a method, apparatus, device, and medium for analyzing multi-target proteins in obesity based on deep learning. Background Technology

[0002] Obesity is a global public health problem. Studies have found that global obesity rates have risen dramatically over the past four decades: childhood obesity rates rose from less than 1% in 1975 to 6%–8% in 2016, adult male obesity rates rose from 3% to 11%, and adult female obesity rates rose from 6% to 15%. Obesity is often closely associated with high mortality rates from cardiovascular disease, type II diabetes, and various cancers, and is a significant risk factor for death and disability. Compared to traditional techniques such as anthropometric measurements, imaging / body composition, and metabolic index testing, molecular phenotyping, while showing better accuracy, cannot directly reflect specific biological processes, and its dynamic response is lagging.

[0003] Proteins are the direct carriers of life activities, providing a more accurate reflection of the body's biological metabolic processes. Proteomics not only provides crucial information on post-transcriptional modifications, protein-protein interactions, and signal transduction, but also offers a basis for drug target discovery and precision medicine. Large-scale proteomic analysis using high-throughput mass spectrometry can screen biomarkers and construct risk prediction models, enabling efficient diagnosis of obesity in individuals and the capture of signals before metabolic disorders, thus achieving early risk prediction and dynamic monitoring of treatment response. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a method, device, equipment and medium for analyzing multi-target proteins of obesity based on deep learning.

[0005] In a first aspect, embodiments of this application provide a method for analyzing multi-target proteins in obesity based on deep learning, including: Biological sample data is acquired and preprocessed to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features; An initial deep learning model for analyzing multi-target proteins in obesity was constructed based on a multi-layer feedforward neural network. The initial deep learning model is trained based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity. Based on the deep learning model used to analyze multi-target proteins in obesity, the expression characteristics of multiple proteins related to obesity were determined.

[0006] In some embodiments, constructing an initial deep learning model for analyzing multi-target proteins in obesity includes: A network hierarchy structure is constructed, comprising an input layer, at least three hidden layers, and an output layer connected in sequence; wherein the input layer, hidden layers, and output layer are all fully connected layer structures. Set the number of neurons in the input layer to be consistent with the number of features of the protein expression features; The number of neurons in each hidden layer is set to decrease sequentially according to a preset feature compression ratio, and a batch normalization component, a leakage correction linear unit nonlinear activation function, and a random deactivation component are sequentially embedded in each hidden layer. Set the number of neurons in the output layer to 1.

[0007] In some embodiments, it also includes: The initial deep learning model is trained using class-weighted cross-entropy as the loss function; wherein the class weights in the class-weighted cross-entropy are calculated based on the original overall label distribution.

[0008] In some embodiments, determining multiple protein expression features related to obesity based on the deep learning model used for analyzing multi-target proteins of obesity includes: Calculate the global importance score corresponding to each protein expression feature; Based on the global importance score, multiple protein expression features associated with obesity were identified.

[0009] In some embodiments, calculating the global importance score corresponding to each protein expression feature includes: For each protein expression feature in the biological sample data, multiple feature subsets are constructed; Based on the multiple feature subsets, calculate the prediction difference when the feature is present and when the feature is absent; Based on the predicted differences, determine the marginal contribution of the protein expression features to the biological sample data; Based on the marginal contribution, the global importance score corresponding to the protein expression feature is determined.

[0010] In some embodiments, determining multiple protein expression features associated with obesity based on the global importance score includes: The protein expression features are sorted in descending order according to the global importance score, and a preset number of the protein expression features with the highest global importance score are identified as obesity-related protein expression features.

[0011] Secondly, embodiments of this application provide a device for analyzing multi-target proteins in obesity based on deep learning, comprising: An acquisition module is used to acquire biological sample data and preprocess the biological sample data to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features; The building block is used to construct an initial deep learning model for analyzing multi-target proteins in obesity based on a multi-layer feedforward neural network. The training module is used to train the initial deep learning model based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity. The analysis module is used to determine the expression characteristics of multiple proteins related to obesity based on the deep learning model used to analyze multi-target proteins in obesity.

[0012] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in embodiments of this application.

[0013] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in embodiments of this application.

[0014] Fifthly, embodiments of this application provide a computer program product, including a computer program, characterized in that, when the computer program is executed by a processor, it implements the method described in embodiments of this application.

[0015] The methods, apparatus, devices, and media for analyzing multi-target proteins in obesity based on deep learning provided in this application utilize large-scale biological sample data for analysis. The sample size is far larger than that of traditional proteomics studies, effectively improving the stability and universality of model analysis and training. An initial deep learning model for analyzing multi-target proteins in obesity is constructed based on a multi-layer feedforward neural network, exhibiting stronger feature extraction capabilities than traditional machine learning models. The use of class-weighted cross-entropy as the loss function effectively addresses the problem of insufficient obesity samples, improving the model's ability to identify a few class entropies. An interpretable global importance scoring algorithm identifies important protein expression features affecting obesity, providing potential targets for clinical research.

[0016] Additional aspects and advantages of the invention 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 the invention. Attached Figure Description

[0017] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 A flowchart illustrating a method for analyzing multiple target proteins in obesity based on deep learning, according to an embodiment of this application, is shown. Figure 2 This is a schematic diagram illustrating the changing trend of the loss value during model training, provided in one embodiment of this application. Figure 3 A schematic diagram of the structure of a device for analyzing multiple obesity-target proteins based on deep learning according to an embodiment of this application is shown; Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device or server according to embodiments of this application is shown. Detailed Implementation

[0018] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] To further illustrate the technical solutions provided in the embodiments of this application, a detailed description is provided below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiments of this application provide method operation instruction steps as shown in the following embodiments or drawings, the method may include more or fewer operation instruction steps based on conventional or non-creative effort. In steps where there is no logically necessary causal relationship, the execution order of these steps is not limited to the execution order provided in the embodiments of this application. In actual processing or when the device executes the method, it may be executed sequentially or in parallel according to the method shown in the embodiments or drawings.

[0021] It should be noted that the acquisition or use of data in the embodiments of this application requires the user's consent. The relevant data can only be obtained after the user's authorization and permission, and the acquisition or use of the data complies with the laws and regulations of the relevant regions.

[0022] Please refer to Figure 1 , Figure 1 This illustration shows a flowchart of a method for analyzing multi-target proteins in obesity based on deep learning, according to an embodiment of this application. Figure 1 As shown, the method includes: Step 101: Obtain biological sample data and preprocess the biological sample data to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features.

[0023] Specifically, biological sample data can be read from a specified data source, such as a biological sample database or data storage server, by calling a data reading interface. This biological sample data includes protein expression data and obesity classification labels. The protein expression data includes various protein expression features. Specifically, a single protein expression dataset includes 2923 protein expression features, which are quantifiable indicators characterizing the state or attribute of a protein in the corresponding object to be analyzed. Each protein expression dataset corresponds to a binary label and is organized and stored in CSV (Comma-Separated Values) format. The binary labels are divided into relevant and irrelevant labels, which can be represented numerically. For example, a relevant label of 1 indicates that the protein expression data is related to obesity, while an irrelevant label of 0 indicates that the protein expression data is not related to obesity.

[0024] Furthermore, a series of standardized and normalized quality optimization and format adaptation operations are performed on the biological sample data to eliminate invalid information in the data, unify the data form, and improve data usability, ultimately forming a preprocessed dataset with a well-structured and high-quality dataset that can be directly used for model training.

[0025] In this embodiment, the specific type of preprocessing operation is not limited. It can be set according to the actual data format type. For example, outlier correction, duplicate data removal, data format normalization, feature dimension filtering and other data processing operations can be configured.

[0026] Step 102: Based on a multi-layer feedforward neural network, construct an initial deep learning model for analyzing multi-target proteins in obesity.

[0027] Among them, the multilayer feedforward neural network can adopt a protein expression network. The initial deep learning model used to analyze multi-target proteins of obesity is a deep learning model used to predict obesity based on biological sample data.

[0028] In some embodiments, constructing an initial deep learning model for analyzing multi-target proteins in obesity includes: constructing a network hierarchy structure comprising an input layer, at least three hidden layers, and an output layer connected in sequence; the input layer, hidden layers, and output layer are all fully connected layer structures; setting the number of neurons in the input layer to be consistent with the number of features of the protein expression features; setting the number of neurons in each hidden layer to decrease sequentially according to a preset feature compression ratio, and embedding a batch normalization component, a leakage correction linear unit nonlinear activation function, and a random deactivation component sequentially in each hidden layer; and setting the number of neurons in the output layer to 1.

[0029] Specifically, a network hierarchy structure is constructed, consisting of an input layer, at least three hidden layers, and an output layer connected sequentially. The network is built in the order of input layer → hidden layer → output layer, with no fewer than three hidden layers. This overcomes the limitation of shallow networks in learning nonlinear relationships and ensures that the model has sufficient depth to learn nonlinear associations. The input layer, hidden layer, and output layer are all fully connected layer structures, with each neuron establishing connections with all neurons in the adjacent layers. This can fully capture the cross-correlation between protein features and adapt to the feature extraction requirements of one-dimensional protein expression data.

[0030] The number of protein expression features corresponds to the number of protein types involved in training. The number of neurons in the input layer corresponds to this number, with each neuron specifically responsible for receiving the expression level data of one protein. This ensures that every protein expression feature (i.e., the expression level of each protein) can be accurately received and processed by the model.

[0031] Furthermore, a feature compression ratio is set, such as an initial hidden layer with 128 neurons, and subsequent layers decreasing by 50% to 64 and 32 neurons respectively. The number of neurons in each hidden layer is configured according to this ratio. Subsequently, in the computation process of each hidden layer, a batch normalization component, a Leaky ReLU activation function (with a slope of 0.1), and a Dropout component (with a probability of 0.2) are added sequentially to ensure that the components work together in the order of normalization → nonlinear activation → regularization. The batch normalization component can stabilize the distribution of input data in each layer and reduce gradient fluctuations; Leaky ReLU is used to introduce nonlinear transformations, which can avoid negative gradient vanishing compared to traditional ReLU and can improve the model's ability to capture complex nonlinear correlations; the Dropout component can randomly shield some neurons, which can prevent the model from over-relying on specific features and effectively suppress overfitting. The collaboration of multiple components ensures a stable and efficient training process.

[0032] Among them, high-dimensional protein expression features contain redundant information, and the design of decreasing number of neurons can gradually filter out invalid features and focus on core information related to obesity.

[0033] In addition, the number of neurons in the output layer corresponds to the output dimension of the model, which is set to 1. This means that the model will ultimately output only a continuous value, namely the prediction result of biological obesity, in order to meet the output requirements of the regression task.

[0034] In the deep learning model with the above architecture, the input layer, hidden layer, and output layer all adopt a fully connected layer structure. This maximizes the capture of global cross-correlation between protein features, avoids the limitations of non-fully connected structures in adapting to one-dimensional protein sequence data, and fits the morphological characteristics of protein expression data. The number of neurons in the input layer is consistent with the number of protein expression features, ensuring that the expression level features of each protein can be individually received and participate in model learning. The number of neurons in the hidden layer decreases sequentially according to a preset ratio, which can gradually compress redundant information from high-dimensional protein expression features and focus on low-dimensional core features strongly correlated with obesity, adapting to the interaction of multiple biological processes in obesity analysis. The model addresses the mutual modeling requirement, balancing model complexity and generalization ability. Batch normalization components stabilize the distribution of input data at each layer, reducing gradient fluctuations and mitigating the problem of divergence during training with high-dimensional data. Leakage correction linear units replace traditional activation functions, effectively preventing negative gradient vanishing and significantly improving the model's ability to capture complex nonlinear relationships between proteins and biological obesity. Random inactivation components, by randomly shielding some neurons, suppress the model's over-reliance on specific features, reducing overfitting risks at the architectural level and enhancing the model's robustness on new data. Setting the output layer to have only one neuron directly adapts to the regression task objective of biological obesity prediction.

[0035] This collaborative architecture enables the model to efficiently process protein expression data from different sources and tissues, and to achieve high prediction accuracy through precise feature extraction and stable training.

[0036] Step 103: Train the initial deep learning model based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity.

[0037] In this embodiment, class-weighted cross-entropy is used as the loss function to train the initial deep learning model. The class weights in the class-weighted cross-entropy are calculated based on the original overall label distribution. Figure 2 The diagram shows the trend of loss value changes during model training.

[0038] Specifically, the following type of weighted cross-entropy is used as the loss function:

[0039]

[0040] in, Weighted cross-entropy for classes, The total number of categories; in this embodiment, K=2. The class weight of category c. For the true label of category c, The deep learning model outputs the class probability of class c. The total number of samples in the training data. This represents the number of samples of category c in the original data.

[0041] Therefore, this application utilizes class-weighted cross-entropy calculated based on the original overall label distribution to fully reflect the degree of data imbalance in the real world, effectively prevent overcompensation, and ensure the generalization ability of the model.

[0042] Specifically, when training the non-insulin-dependent diabetes analysis model built based on deep learning, the Adam optimizer was also used with an initial learning rate of 0.001, and end-to-end optimization was performed in combination with the class-weighted cross-entropy loss function to improve the model's classification performance under imbalanced data.

[0043] In some examples, this also includes: reliability verification of the trained deep learning model.

[0044] Specifically, the reliability verification of the trained deep learning model includes: calculating at least one evaluation metric of the deep learning model; if at least one evaluation metric satisfies the reliability verification conditions, then the analysis model is determined to satisfy the reliability verification.

[0045] Specifically, the evaluation metrics can be computational accuracy, AUC (Area Under the ROC Curve), and loss value. If the computational accuracy exceeds the preset computational accuracy, the AUC exceeds the preset AUC, and the loss value is lower than the preset loss value, then the reliability verification is considered to be satisfied.

[0046] AUC reflects the classification ability of the deep learning model. Accuracy reflects the classification accuracy, and loss reflects the prediction error. Lower values ​​indicate a smaller difference between the model's output probability distribution and the true labels, resulting in a better fit. Preset accuracy can be set to any value above 50%, AUC to any value above 0.5, and loss to any value below 1. Higher preset accuracy and AUC place higher demands on the deep learning model and result in better model performance. Conversely, lower preset loss also places higher demands on the deep learning model and results in better model performance.

[0047] For example, 53,013 biological sample data are input into the data input module. These 53,013 biological sample data are used as raw data and undergo median imputation, standardization, and data balancing in the preprocessing module to obtain preprocessed biological sample data. This preprocessed biological sample data is then used as the preprocessed sample data and input into an initial deep learning model for analyzing multi-target proteins in obesity, constructed based on a multi-layer feedforward neural network. The model is then trained. Specifically, during training, the training module uses Adam optimization, class-weighted cross-loss, and an early stopping mechanism to update the model parameters. The evaluation module is then used to assess whether the trained model can be used as a deep learning model for analyzing multi-target proteins in obesity. If it can, the deep learning model is used to identify key proteins.

[0048] Step 104: Based on a deep learning model used to analyze multi-target proteins in obesity, determine the expression characteristics of multiple proteins associated with obesity.

[0049] Specifically, by calculating the global importance score corresponding to each protein expression feature, and then determining multiple protein expression features related to obesity based on the global importance score.

[0050] SHAP (Shapley Additive exPlanations) is a game theory-based model interpretation method used to calculate the contribution of each feature to the model's predictions. Its core idea is to use Shapley values ​​to fairly distribute the contribution of each feature in the prediction results. Based on this, this application uses SHAP to analyze protein expression features related to obesity.

[0051] In one feasible embodiment, multiple feature subsets are constructed for each protein expression feature in the biological sample data; based on the multiple feature subsets, the prediction difference with and without the feature is calculated respectively; based on the prediction difference, the marginal contribution of the protein expression feature to the biological sample data is determined; based on the marginal contribution, the global importance score corresponding to the protein expression feature is determined.

[0052] In a preferred embodiment, a background set is constructed before constructing multiple feature subsets. The background set serves to provide reference values ​​for subsequent feature "missing" analysis.

[0053] For example, the global importance score is calculated using the following formula:

[0054] in, The global importance score for the expression feature of the j-th protein. denoted as the marginal contribution of the j-th protein expression feature to the x-th biological sample data, where N is the total number of protein expression features contained in the biological sample data.

[0055] Furthermore, after obtaining the global importance scores corresponding to all protein expression features, the protein expression features are sorted in descending order according to the global importance scores, and a preset number of protein expression features with the highest global importance scores are identified as obesity-related protein expression features.

[0056] Specifically, the ten proteins highly associated with obesity were identified using the methods proposed in the embodiments of this application: LEP, DKKL1, FABP4, ADAMTS16, FOLR3, GFAP, CSF2RA, KCNC4, LTO1, and CCL15. Among them, leptin (LEP), secreted by adipose tissue, is a hormone that regulates appetite, neuroendocrine function, and energy balance; its abnormalities are closely related to various metabolic syndromes, especially obesity. FABP4 is mainly expressed in adipocytes and macrophages and participates in insulin resistance and atherosclerosis; studies have shown that elevated plasma FABP4 levels are significantly associated with obesity, diabetes, hypertension, and cardiovascular events. These highly important proteins are all involved in metabolic, inflammatory, and immune pathways, confirming the view that proteins act as mediators of obesity pathology.

[0057] In summary, the method for analyzing multiple target proteins in obesity based on deep learning provided in this application utilizes large-scale biological sample data for analysis, with a sample size far exceeding that of traditional proteomics studies, effectively improving the stability and universality of model analysis and training. The initial deep learning model for analyzing multiple target proteins in obesity, constructed based on a multi-layer feedforward neural network, exhibits stronger feature extraction capabilities than traditional machine learning models. The use of class-weighted cross-entropy as the loss function effectively addresses the problem of insufficient obesity samples, improving the model's ability to identify a few class entropies. Through an interpretable global importance scoring algorithm, the method identifies important protein expression features affecting obesity, providing potential targets for clinical research.

[0058] It should be noted that although the operation of the method of the present invention is described in a specific order in the accompanying drawings, this does not require or imply that the operations must be performed in that specific order, or that all the operations shown must be performed in order to achieve the desired result.

[0059] Figure 3 A schematic diagram of the structure of a device for analyzing multi-target proteins of obesity based on deep learning, provided in an embodiment of this application, is shown.

[0060] like Figure 3 As shown, the device 10 for analyzing multi-target proteins in obesity based on deep learning includes: The acquisition module 11 is used to acquire biological sample data and preprocess the biological sample data to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features; Module 12 is used to build an initial deep learning model for analyzing multi-target proteins in obesity based on a multi-layer feedforward neural network. Training module 13 is used to train the initial deep learning model based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity. Analysis module 14 is used to determine the expression characteristics of multiple proteins related to obesity based on the deep learning model used to analyze multi-target proteins of obesity.

[0061] In some embodiments, the construction module 12 is specifically used for: A network hierarchy structure is constructed, comprising an input layer, at least three hidden layers, and an output layer connected in sequence; wherein the input layer, hidden layers, and output layer are all fully connected layer structures. Set the number of neurons in the input layer to be consistent with the number of features of the protein expression features; The number of neurons in each hidden layer is set to decrease sequentially according to a preset feature compression ratio, and a batch normalization component, a leakage correction linear unit nonlinear activation function, and a random deactivation component are sequentially embedded in each hidden layer. Set the number of neurons in the output layer to 1.

[0062] In some embodiments, the training module 13 is specifically used for: The initial deep learning model is trained using class-weighted cross-entropy as the loss function; wherein the class weights in the class-weighted cross-entropy are calculated based on the original overall label distribution.

[0063] In some embodiments, the analysis module 14 is specifically used for: Calculate the global importance score corresponding to each protein expression feature; Based on the global importance score, multiple protein expression features associated with obesity were identified.

[0064] In some embodiments, the analysis module 14 is specifically used for: For each protein expression feature in the biological sample data, multiple feature subsets are constructed; Based on the multiple feature subsets, calculate the prediction difference when the feature is present and when the feature is absent; Based on the predicted differences, determine the marginal contribution of the protein expression features to the biological sample data; Based on the marginal contribution, the global importance score corresponding to the protein expression feature is determined.

[0065] In some embodiments, the analysis module 14 is specifically used for: The protein expression features are sorted in descending order according to the global importance score, and a preset number of the protein expression features with the highest global importance score are identified as obesity-related protein expression features.

[0066] It should be understood that the modules or modules described in the device 10 for analyzing multi-target proteins in obesity based on deep learning are similar to those in the reference device. Figure 1 The steps in the described method correspond accordingly. Therefore, the operations and features described above for the method are also applicable to the device 10 for analyzing multi-target proteins of obesity based on deep learning and its included modules, and will not be repeated here. The device 10 for analyzing multi-target proteins of obesity based on deep learning can be pre-implemented in the browser or other secure applications of an electronic device, or can be loaded into the browser or its secure applications of an electronic device through download or other means. The corresponding modules in the device 10 for analyzing multi-target proteins of obesity based on deep learning can cooperate with the modules in the electronic device to implement the solutions of the embodiments of this application.

[0067] The division of modules or units mentioned in the detailed description above is not mandatory. In fact, according to the embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0068] The following is for reference. Figure 4 , Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the embodiments of this application is shown. like Figure 4 As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 402 or programs loaded from storage section 408 into random access memory (RAM) 403. RAM 403 also stores various programs and data required for the system's operating instructions. CPU 401, ROM 402, and RAM 403 are interconnected via bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0069] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0070] Specifically, according to embodiments of this application, the flowchart above refers to... Figure 2 The described process can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such an embodiment, the computer program contains program code for performing the methods shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the functions defined in the system of this application.

[0071] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0072] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operational instructions of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two connected blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified functions or operational instructions, or using a combination of dedicated hardware and computer instructions.

[0073] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be housed in a processor; for example, a processor can be described as including an acquisition module, a construction module, a training module, and an analysis module. The names of these units or modules do not necessarily limit the specific unit or module itself. For example, an acquisition module can also be described as "acquiring biological sample data and preprocessing the biological sample data to obtain preprocessed biological sample data; the preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features."

[0074] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments, or may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium stores one or more programs that, when used by one or more processors, execute the method described in this application for analyzing multiple obesity-related proteins based on deep learning.

[0075] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for analyzing multi-target proteins in obesity based on deep learning, characterized in that, include: Biological sample data is acquired and preprocessed to obtain preprocessed biological sample data. The preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features. An initial deep learning model for analyzing multi-target proteins in obesity was constructed based on a multi-layer feedforward neural network. The initial deep learning model is trained based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity. Based on the deep learning model used to analyze multi-target proteins in obesity, the expression characteristics of multiple proteins related to obesity were determined.

2. The method for analyzing multi-target proteins in obesity based on deep learning according to claim 1, characterized in that, The construction of the initial deep learning model for analyzing multi-target proteins in obesity includes: A network hierarchy structure is constructed, comprising an input layer, at least three hidden layers, and an output layer connected in sequence; wherein the input layer, hidden layers, and output layer are all fully connected layer structures. Set the number of neurons in the input layer to be consistent with the number of features of the protein expression features; The number of neurons in each hidden layer is set to decrease sequentially according to a preset feature compression ratio, and a batch normalization component, a leakage correction linear unit nonlinear activation function, and a random deactivation component are sequentially embedded in each hidden layer. Set the number of neurons in the output layer to 1.

3. The method for analyzing multi-target proteins in obesity based on deep learning according to claim 1 or 2, characterized in that, Also includes: The initial deep learning model is trained using class-weighted cross-entropy as the loss function; wherein the class weights in the class-weighted cross-entropy are calculated based on the original overall label distribution.

4. The method for analyzing multi-target proteins in obesity based on deep learning according to claim 1, characterized in that, The deep learning model used for analyzing multi-target proteins related to obesity determines the expression features of multiple proteins associated with obesity, including: Calculate the global importance score corresponding to each protein expression feature; Based on the global importance score, multiple protein expression features associated with obesity were identified.

5. The method for analyzing multi-target proteins in obesity based on deep learning according to claim 4, characterized in that, The calculation of the global importance score corresponding to each protein expression feature includes: For each protein expression feature in the biological sample data, multiple feature subsets are constructed; Based on the multiple feature subsets, calculate the prediction difference when the feature is present and when the feature is absent; Based on the predicted differences, determine the marginal contribution of the protein expression features to the biological sample data; Based on the marginal contribution, the global importance score corresponding to the protein expression feature is determined.

6. The method for analyzing multi-target proteins in obesity based on deep learning according to claim 4, characterized in that, The determination of multiple protein expression features related to obesity based on the global importance score includes: The protein expression features are sorted in descending order according to the global importance score, and a preset number of the protein expression features with the highest global importance score are identified as obesity-related protein expression features.

7. A device for analyzing multi-target proteins in obesity based on deep learning, characterized in that, include: The acquisition module is used to acquire biological sample data and preprocess the biological sample data to obtain preprocessed biological sample data. The preprocessed biological sample data includes protein expression data and obesity classification labels corresponding to the protein expression data; the protein expression data includes multiple protein expression features. The building block is used to construct an initial deep learning model for analyzing multi-target proteins in obesity based on a multi-layer feedforward neural network. The training module is used to train the initial deep learning model based on the preprocessed biological sample data to obtain a deep learning model for analyzing multi-target proteins in obesity. The analysis module is used to determine the expression characteristics of multiple proteins related to obesity based on the deep learning model used to analyze multi-target proteins in obesity.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for analyzing multiple target proteins of obesity based on deep learning as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the method for analyzing multi-target proteins of obesity based on deep learning as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the method for analyzing multi-target proteins of obesity based on deep learning, as described in any one of claims 1-6.

Images