Drug effect prediction model construction method and device, electronic equipment and storage medium

Abstract

The invention discloses a drug effect prediction model construction method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining physiological feature data of a plurality of sample objects and real pharmacokinetic data and real drug effect data of a target drug relative to each sample object; according to the multiple physiological feature data, the real pharmacokinetic data and the real pharmacodynamic data, constructing a pharmacodynamic parameter prediction model; performing data simulation on the plurality of physiological feature data to obtain physiological feature data of a plurality of simulation sample objects; inputting the physiological characteristic data of the simulation sample object into a pharmacodynamic parameter prediction model to obtain simulation pharmacodynamic data; and training according to the physiological feature data and the simulated drug effect data of the plurality of simulated sample objects to obtain a target drug effect prediction model, and determining a drug application object of the target drug based on the target drug effect prediction model. According to the technical scheme, the effect of accurately determining the ideal drug effect object before the clinical test is carried out is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of the medical industry, and in particular to a method, device, electronic equipment and storage medium for constructing a drug efficacy prediction model. Background Art

[0002] As we all know, the research and development cycle of new drugs is long, the investment is huge, and the risk is high. It has always been a job that requires "slow work and fine work" and "one wrong move will lead to a complete failure."

[0003] At present, the traditional way of drug development is to conduct clinical trials on drugs, and then determine the drug development results based on the results of clinical trials.

[0004] However, drug development based on clinical trials may result in clinical trial failure due to lack of drug efficacy data or inaccurate subject selection, which in turn affects drug development efficiency and increases drug development costs. Summary of the invention

[0005] The present invention provides a method, device, electronic device and storage medium for constructing a drug efficacy prediction model, so as to accurately simulate drug efficacy data before conducting clinical trials, thereby accurately determining the effect of an ideal drug efficacy object, thereby achieving the effect of improving the success rate and efficiency of drug development on the basis of reducing drug development costs.

[0006] According to one aspect of the present invention, a method for constructing a drug efficacy prediction model is provided, the method comprising:

[0007] Acquiring physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of a target drug relative to each of the sample objects;

[0008] Constructing a drug efficacy parameter prediction model corresponding to the target drug according to the plurality of physiological characteristic data, the real pharmacokinetic data and the real drug efficacy data;

[0009] Performing data simulation on the plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects;

[0010] For each of the simulated sample objects, inputting the physiological characteristic data corresponding to the simulated sample object into the drug efficacy parameter prediction model to obtain the simulated drug efficacy data of the target drug relative to the simulated sample object;

[0011] A plurality of training samples are constructed according to the physiological characteristic data corresponding to the plurality of simulated sample objects and the simulated pharmacodynamic data of the target drug relative to the simulated sample objects, and a target pharmacodynamic prediction model corresponding to the target drug is obtained based on the training of the plurality of training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested.

[0012] According to another aspect of the present invention, a device for constructing a drug efficacy prediction model is provided, the device comprising:

[0013] A data acquisition module, used to acquire physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each of the sample objects;

[0014] A model building module, used to build a pharmacodynamic parameter prediction model corresponding to the target drug according to the plurality of physiological characteristic data, the real pharmacokinetic data and the real pharmacodynamic data;

[0015] A data simulation module, used to perform data simulation on the plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects;

[0016] A data prediction module, for inputting the physiological characteristic data corresponding to each simulated sample object into the drug efficacy parameter prediction model to obtain the simulated drug efficacy data of the target drug relative to the simulated sample object;

[0017] A model training module is used to construct a plurality of training samples based on the plurality of physiological characteristic data, the real pharmacodynamic data, the physiological characteristic data corresponding to the plurality of simulated sample objects, and the simulated pharmacodynamic data of the target drug relative to the simulated sample objects, and to obtain a target pharmacodynamic prediction model corresponding to the target drug based on the plurality of training samples, so as to determine a drug applicable object corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested.

[0018] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0019] at least one processor; and

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for constructing a drug efficacy prediction model described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for constructing a drug efficacy prediction model described in any embodiment of the present invention when executed.

[0023] The technical solution of the embodiment of the present invention solves the problem in the related art that clinical trials fail due to lack of drug efficacy data or inaccurate selection of subjects, thereby affecting drug development efficiency and increasing drug development costs. It achieves accurate simulation of drug efficacy data before conducting clinical trials, and then accurately determines the effect of ideal drug efficacy objects, thereby achieving the effect of improving drug development success rate and development efficiency on the basis of reducing drug development costs.

[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0026] Figure 1 is a flow chart of a method for constructing a drug efficacy prediction model provided in accordance with Embodiment 1 of the present invention;

[0027] Figure 2 is a flow chart of a method for constructing a drug efficacy prediction model provided in accordance with Embodiment 2 of the present invention;

[0028] Figure 3 is a schematic diagram of the structure of a drug efficacy prediction model building device provided according to Embodiment 3 of the present invention;

[0029] Figure 4 It is a structural schematic diagram of an electronic device for implementing the method for constructing a drug efficacy prediction model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Embodiment 1

[0033] Figure 1 This is a flow chart of a method for constructing a drug efficacy prediction model provided by the first embodiment of the present invention. This embodiment can be applied to predicting the drug efficacy data of a target drug based on the physiological characteristic data of the subject to be tested, and determining an ideal drug efficacy object based on the drug efficacy data. This method can be executed by a drug efficacy prediction model construction device, which can be implemented in the form of hardware and / or software, and can be configured in a terminal and / or server. Figure 1 As shown, the method includes:

[0034] S110 , obtaining physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each sample object.

[0035] Among them, the sample object may be an object that has received the target drug and has recorded the efficacy data. Physiological characteristic data can be understood as special characteristics or attributes that characterize the corresponding object at the biological level. Physiological characteristic data may include characteristic values ​​corresponding to multiple physiological characteristics. Optionally, physiological characteristics may include but are not limited to height, weight, age, blood routine, blood biochemistry, body temperature, pulse, systolic blood pressure, and diastolic blood pressure. The target drug may be a drug for which drug applicable characteristics are to be predicted. The target drug may be any drug. Real pharmacokinetic data may be understood as parameters that characterize the absorption, distribution, metabolism, and excretion process of the corresponding drug in the subject. Optionally, real pharmacokinetic data may include but are not limited to drug absorption parameters (such as permeability, absorption rate, etc.), drug distribution parameters (such as plasma protein binding rate, etc.), drug metabolism parameters (such as drug plasma concentration, etc.), drug excretion parameters, and area under the drug-time curve (Area Under the plasma concentration-time Curve, AUC), etc. Real efficacy data may be efficacy data detected in real time during the process of the target drug acting on the sample object.

[0036] In this embodiment, there may be multiple ways to obtain physiological characteristic data. Optionally, when multiple sample objects are determined, for each sample object, pre-stored characteristic values ​​corresponding to multiple physiological characteristics corresponding to the sample object may be retrieved from a pre-built database. Then, the physiological characteristic data corresponding to the sample object may be obtained; or, for each sample object, the sample object may be detected by a corresponding characteristic detection device to obtain the characteristic value corresponding to the sample object under the physiological characteristic. Then, the physiological characteristic data corresponding to the sample object may be obtained; or, the physiological characteristic data corresponding to the sample object may be obtained by other means, which is not specifically limited in this embodiment.

[0037] It should be noted that there may be multiple ways to obtain the real pharmacokinetic data and real pharmacodynamic data of the target drug relative to the sample object, which may optionally include retrieving from a target database; or, subjecting the target drug to the sample object, and then performing real-time detection on the sample object to obtain corresponding detection data, and determining the real pharmacokinetic data and real pharmacodynamic data based on the detection data; or, it may be obtained through other methods, which are not specifically limited in this embodiment.

[0038] In this embodiment, in order to construct a drug efficacy prediction model, the physiological characteristic data corresponding to the sample objects used for model construction and the real pharmacokinetic data and real drug efficacy data of the target drug relative to each sample object can be obtained first. Then, the corresponding drug efficacy prediction model can be constructed based on the acquired sample data.

[0039] S120, constructing a pharmacodynamic parameter prediction model corresponding to the target drug based on multiple physiological characteristic data, real pharmacokinetic data and real pharmacodynamic data.

[0040] Among them, the pharmacodynamic parameter prediction model can be understood as a model that takes physiological characteristic data as input objects to predict the pharmacodynamic data of the target drug based on the physiological characteristic data. The pharmacodynamic parameter prediction model can be a model of a mathematical function structure, that is, a model constructed based on a data function relationship. It should be noted that the pharmacodynamic parameter prediction model can be a model for enhancing training data. In other words, the advantage of constructing a pharmacodynamic parameter prediction model is that when there is less training data for training the target pharmacodynamic prediction model, the physiological characteristic data can be simulated to obtain a large amount of physiological characteristic data, and then, the pharmacodynamic data corresponding to each physiological characteristic data can be predicted based on the pharmacodynamic parameter prediction model to achieve the effect of training data enhancement.

[0041] In this embodiment, physiological characteristic data corresponding to multiple sample objects, real pharmacokinetic data of the target drug relative to the sample objects, and real pharmacodynamic data of the target drug relative to the sample objects can be obtained. Further, the mapping relationship between the physiological characteristic data and the pharmacokinetic data can be analyzed based on the physiological characteristic data and the real pharmacokinetic data corresponding to the multiple sample objects, and the mapping relationship between the pharmacokinetic data and the pharmacodynamic data can be analyzed based on the real pharmacokinetic data and the real pharmacodynamic data, and then, the pharmacodynamic parameter prediction model can be constructed based on the two determined mapping relationships.

[0042] Optionally, a pharmacodynamic parameter prediction model corresponding to the target drug is constructed based on multiple physiological characteristic data, real pharmacokinetic data and real pharmacodynamic data, including: determining a mapping relationship between multiple physiological characteristic data and multiple real pharmacokinetic data, and using the mapping relationship as a pharmacokinetic parameter prediction sub-model; determining at least one candidate pharmacokinetic-pharmacodynamic binding sub-model corresponding to the target drug based on the drug action mechanism corresponding to the target drug, and determining a pharmacokinetic-pharmacodynamic binding sub-model based on the real pharmacokinetic data, the real pharmacodynamic data and at least one candidate pharmacokinetic-pharmacodynamic binding sub-model; and connecting the pharmacokinetic parameter prediction sub-model and the pharmacokinetic-pharmacodynamic binding sub-model to obtain a pharmacodynamic parameter prediction model.

[0043] Among them, the pharmacokinetic parameter prediction submodel can be a model that takes physiological characteristic data as an input object to predict pharmacokinetic parameters based on physiological characteristic data. The mechanism of action of a drug refers to information about where, how and why a drug works. The candidate pharmacokinetic-pharmacodynamic binding submodel can be understood as a pharmacokinetic-pharmacodynamic binding submodel to be selected. The pharmacokinetic-pharmacodynamic binding submodel (Pharmacokinetic-Pharmacodynamic, PK-PD) can be understood as a model that takes pharmacokinetic data as an input object to predict pharmacodynamic data based on pharmacokinetic data. Exemplarily, if the drug action mechanism of the target drug is a physiological effect indirectly produced by affecting endogenous substances, the candidate pharmacokinetic-pharmacodynamic binding submodel can be a flip model.

[0044] In this embodiment, the physiological characteristic data and the real pharmacokinetic data corresponding to each sample object can be used as a set of data, and multiple sets of data can be obtained. After that, data analysis can be performed on the multiple sets of data to determine the mapping relationship between the physiological characteristic data and the real pharmacokinetic data according to the data analysis results. Furthermore, a model can be constructed based on the mapping relationship, and the constructed model can be used as a pharmacokinetic parameter prediction submodel. Further, the drug action mechanism corresponding to the target drug can be obtained, and at least one candidate pharmacokinetic-pharmacodynamic binding submodel corresponding to the target drug can be determined based on the drug action mechanism. Afterwards, the pharmacokinetic-pharmacodynamic binding submodel can be determined based on the real pharmacokinetic data of the target drug relative to the sample object, the real pharmacodynamic data corresponding to the target drug, and at least one candidate pharmacokinetic-pharmacodynamic binding submodel.

[0045] Optionally, a pharmacokinetic-pharmacodynamic binding submodel is determined based on real pharmacokinetic data, real pharmacodynamic data and at least one candidate pharmacokinetic-pharmacodynamic binding submodel, including: for each candidate pharmacokinetic-pharmacodynamic binding submodel, inputting the real pharmacokinetic data into the candidate pharmacokinetic-pharmacodynamic binding submodel to obtain pharmacodynamic prediction data corresponding to the target drug; processing the pharmacodynamic prediction data and the real pharmacodynamic data corresponding to the target drug according to at least one preset evaluation indicator to obtain an evaluation result corresponding to the candidate pharmacokinetic-pharmacodynamic binding submodel under each preset evaluation indicator; and determining a pharmacokinetic-pharmacodynamic binding submodel from at least one candidate pharmacokinetic-pharmacodynamic binding submodel according to at least one evaluation result.

[0046] The preset evaluation index is used to evaluate the model prediction performance of the candidate pharmacokinetic-pharmacodynamic binding submodel; optionally, the preset evaluation index may include goodness of fit (GOF) and area under the plasma concentration-time curve (AUC). The evaluation result corresponding to the goodness of fit can be obtained by calculating the Pearson determination coefficient (R 2 The evaluation result corresponding to the area under the drug-time curve can be determined by calculating the mean absolute error (MAE) and the mean prediction error (MFE).

[0047] As an optional implementation of the first embodiment of this embodiment, for each candidate pharmacokinetic-pharmacodynamic combination submodel, the real pharmacokinetic data of the target drug relative to the sample object can be input into the candidate pharmacokinetic-pharmacodynamic combination submodel, and the pharmacodynamic prediction data corresponding to the target drug can be obtained. Further, the pharmacodynamic prediction data and the real pharmacodynamic data can be processed according to at least one preset evaluation index, and an evaluation result corresponding to each preset evaluation index can be obtained. Further, at least one evaluation result can be compared to determine the evaluation result with the smallest parameter value. Then, the candidate pharmacokinetic-pharmacodynamic combination submodel corresponding to the evaluation result can be used as the pharmacokinetic-pharmacodynamic combination submodel corresponding to the target drug. Further, the pharmacokinetic parameter prediction submodel and the pharmacokinetic-pharmacodynamic combination submodel can be connected together to use the output parameters of the pharmacokinetic parameter prediction submodel as the input parameters of the pharmacokinetic-pharmacodynamic combination submodel. Then, the model obtained after the connection can be used as a pharmacodynamic parameter prediction model.

[0048] S130 , performing data simulation on a plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects.

[0049] In this embodiment, in order to enhance the training data, the physiological characteristic data corresponding to the sample object can be simulated according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to multiple simulated sample objects. The data simulation algorithm can be understood as an algorithm for simulating data to expand the sample data. Optionally, the data simulation algorithm can be a Monte Carlo simulation algorithm. The simulated sample object can be understood as a sample object simulated based on the data simulation algorithm.

[0050] As an optional implementation manner of this embodiment, data simulation may be performed on the physiological characteristic data corresponding to the sample object according to a predetermined data simulation algorithm, and physiological characteristic data corresponding to a plurality of simulated sample objects may be obtained.

[0051] S140 . For each simulated sample object, input the physiological characteristic data corresponding to the simulated sample object into a drug efficacy parameter prediction model to obtain simulated drug efficacy data of the target drug relative to the simulated sample object.

[0052] In this embodiment, for each simulated sample object, the physiological characteristic data corresponding to the simulated sample object can be input into the drug efficacy parameter prediction model. Furthermore, the physiological characteristic data can be processed based on the drug efficacy parameter prediction model, and the data output by the model can be used as the simulated drug efficacy data of the target drug relative to the simulated sample object.

[0053] S150. Construct a plurality of training samples according to the physiological characteristic data corresponding to a plurality of simulated sample objects and the simulated efficacy data of the target drug relative to the simulated sample objects, and obtain a target efficacy prediction model corresponding to the target drug based on the training of the plurality of training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target efficacy prediction model and the physiological characteristic data corresponding to the object to be tested.

[0054] Among them, the target efficacy prediction model can be a neural network model that takes physiological characteristic data as an input object to predict the efficacy data of the target drug relative to the object to be tested based on the physiological characteristic data. The target efficacy prediction model is trained based on the physiological characteristic data corresponding to the sample object and the real efficacy data of the target drug relative to the sample object. The target efficacy prediction model can be a neural network model of any model structure. Optionally, the target efficacy prediction model can be a support vector machine (SVM), a random forest (RF) model, a gradient boosting machine (GBM) model, and a deep neural network (DNN) model. It should be noted that the target efficacy prediction model corresponds to the target drug, that is, each target drug will correspond to a different target efficacy prediction model. The object to be tested can be an object that receives the drug. The object to be tested can be any type of organism. Optionally, the types of objects to be tested can include but are not limited to rodents, dogs, rabbits, and fish. The drug applicable object can be understood as a drug action object that enables the drug to exert the desired ideal efficacy.

[0055] In this embodiment, multiple training samples can be constructed based on the physiological characteristic data corresponding to multiple simulated sample objects and the simulated efficacy data of the target drug relative to the simulated sample objects. Further, a target efficacy prediction model corresponding to the target drug can be obtained by training based on the multiple training samples.

[0056] Optionally, a target pharmacodynamic prediction model corresponding to the target drug is obtained based on training of multiple training samples, including: for multiple training samples, inputting the physiological characteristic data corresponding to the simulated sample objects in the training samples into the pharmacodynamic prediction model to be trained, to obtain pharmacodynamic prediction data corresponding to the training samples; determining a loss value based on the simulated pharmacodynamic data and the pharmacodynamic prediction data in the training samples, and correcting the model parameters in the pharmacodynamic prediction model to be trained based on the loss value, until the loss function in the pharmacodynamic prediction model to be trained converges, to obtain the target pharmacodynamic prediction model corresponding to the target drug.

[0057] In this embodiment, for multiple training samples, the physiological characteristic data corresponding to the simulated sample objects in the training samples can be input into the pharmacodynamic prediction model to be trained to obtain the pharmacodynamic prediction data corresponding to the training samples. Further, the pharmacodynamic prediction data can be compared with the real pharmacodynamic data in the training samples to determine the loss value. Furthermore, the model parameters in the pharmacodynamic prediction model to be trained can be corrected according to the loss value. Afterwards, the training error of the loss function in the pharmacodynamic prediction model to be trained, that is, the loss function, can be used as a condition for detecting whether the current loss function has reached convergence, for example, whether the training error is less than the preset error or whether the error change trend tends to be stable, or whether the current model iteration number is equal to the preset number of times. If the convergence condition is detected, for example, the training error of the loss function is less than the preset error or the error change tends to be stable, it indicates that the training of the pharmacodynamic prediction model to be trained is completed, and the iterative training can be stopped at this time. If it is detected that the convergence condition is not reached at present, the training sample can be further obtained to train the pharmacodynamic prediction model to be trained until the training error of the loss function is within the preset range. When the training error of the loss function reaches convergence, the drug efficacy prediction model to be trained obtained by the current training can be used as the target drug efficacy prediction model.

[0058] In this embodiment, when the target drug efficacy prediction model is obtained, the physiological characteristic data corresponding to the object to be tested can be obtained. Furthermore, the physiological characteristic data corresponding to the object to be tested can be processed based on the target drug efficacy prediction model corresponding to the target drug to determine the drug applicable object corresponding to the target drug.

[0059] The technical solution of the embodiment of the present invention obtains physiological characteristic data corresponding to multiple sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each sample object, and then, according to the multiple physiological characteristic data, the real pharmacokinetic data and the real pharmacodynamic data, constructs a pharmacodynamic parameter prediction model corresponding to the target drug, and then, according to a predetermined data simulation algorithm, performs data simulation on the multiple physiological characteristic data to obtain physiological characteristic data corresponding to multiple simulated sample objects, and then, for each simulated sample object, inputs the physiological characteristic data corresponding to the simulated sample object into the pharmacodynamic parameter prediction model to obtain simulated pharmacodynamic data of the target drug relative to the simulated sample object, and then, according to the multiple simulated sample objects, Corresponding physiological characteristic data and simulated pharmacodynamic data of the target drug relative to the simulated sample object are used to construct multiple training samples, and a target pharmacodynamic prediction model corresponding to the target drug is obtained based on the multiple training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested. This solves the problem in the related technology that clinical trials fail due to lack of drug efficacy data or inaccurate selection of subjects, thereby affecting drug development efficiency and increasing drug development costs, and achieves accurate simulation of pharmacodynamic data before conducting clinical trials, thereby accurately determining the effect of ideal pharmacodynamic objects, thereby achieving the effect of improving drug development success rate and development efficiency on the basis of reducing drug development costs.

[0060] Embodiment 2

[0061] Figure 2 This is a flow chart of a method for constructing a drug efficacy prediction model provided by the second embodiment of the present invention. On the basis of the above embodiments, when the target drug efficacy prediction model is obtained, the drug applicable object corresponding to the target drug can be determined based on the target drug efficacy prediction model and the physiological characteristic data corresponding to the object to be tested. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same or similar to the above embodiments are not repeated here.

[0062] like Figure 2 As shown, the method includes:

[0063] S210, acquiring physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each sample object.

[0064] S220. Construct a pharmacodynamic parameter prediction model corresponding to the target drug based on multiple physiological characteristic data, real pharmacokinetic data, and real pharmacodynamic data.

[0065] S230 , performing data simulation on a plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects.

[0066] S240 . For each simulated sample object, input the physiological characteristic data corresponding to the simulated sample object into a drug efficacy parameter prediction model to obtain simulated drug efficacy data of the target drug relative to the simulated sample object.

[0067] S250. Construct a plurality of training samples according to the physiological characteristic data corresponding to a plurality of simulated sample objects and the simulated efficacy data of the target drug relative to the simulated sample objects, and obtain a target efficacy prediction model corresponding to the target drug based on the training of the plurality of training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target efficacy prediction model and the physiological characteristic data corresponding to the object to be tested.

[0068] S260: Acquire physiological characteristic data corresponding to a plurality of objects to be tested, wherein the physiological characteristic data includes characteristic values ​​corresponding to a plurality of physiological characteristics.

[0069] Among them, the object to be tested can be an object receiving a drug. The object to be tested can be any type of organism. Optionally, the type of the object to be tested can include but is not limited to rodents, dogs, rabbits, and fish. Physiological characteristic data can be understood as special characteristics or attributes that characterize the corresponding object at the biological level. Physiological characteristic data may include characteristic values ​​corresponding to multiple physiological characteristics. Optionally, physiological characteristics may include but are not limited to height, weight, age, blood routine, blood biochemistry, body temperature, pulse, systolic blood pressure, and diastolic blood pressure.

[0070] In this embodiment, there may be multiple ways to obtain physiological characteristic data. Optionally, when multiple objects to be tested are determined, for each object to be tested, characteristic values ​​corresponding to multiple physiological characteristics corresponding to the object to be tested that are pre-stored can be retrieved from a pre-constructed database. Then, the physiological characteristic data corresponding to the object to be tested can be obtained; or, for each object to be tested, the object to be tested can be detected by a corresponding characteristic detection device to obtain the characteristic value corresponding to the object to be tested under the physiological characteristic. Then, the physiological characteristic data corresponding to the object to be tested can be obtained; or, the physiological characteristic data corresponding to the object to be tested can also be obtained by other means, which is not specifically limited in this embodiment.

[0071] S270 , for each subject to be tested, the physiological characteristic data corresponding to the subject to be tested is processed according to a pre-trained target drug efficacy prediction model corresponding to the target drug, to obtain first drug efficacy prediction data of the target drug relative to the subject to be tested.

[0072] Among them, the target drug can be a drug for which drug applicable characteristics are to be predicted. The target drug can be any drug. The target drug efficacy prediction model can be a neural network model that takes physiological characteristic data as an input object to predict the drug efficacy data of the target drug relative to the object to be tested based on the physiological characteristic data. The target drug efficacy prediction model is trained based on the physiological characteristic data corresponding to the sample object and the real drug efficacy data of the target drug relative to the sample object. The target drug efficacy prediction model can be a neural network model of any model structure. Optionally, the target drug efficacy prediction model can be a support vector machine (SVM), a random forest (RF) model, a gradient boosting machine (GBM) model, and a deep neural network (DNN) model. It should be noted that the target drug efficacy prediction model corresponds to the target drug, that is, each target drug will correspond to a different target drug efficacy prediction model. The first drug efficacy prediction data can be based on model prediction, and is used to characterize the drug efficacy of the target drug on the object to be tested. The first drug efficacy prediction data may include an index value corresponding to at least one drug efficacy index. It should be noted that the drug efficacy index included in the first drug efficacy prediction data corresponds to the target drug, that is, different target drugs may correspond to different drug efficacy indexes.

[0073] In this embodiment, when the physiological characteristic data corresponding to multiple objects to be tested are obtained, the target drug can be determined, and the pre-trained target drug efficacy prediction model corresponding to the target drug can be retrieved. Further, for each object to be tested, the physiological characteristic data corresponding to the object to be tested can be input into the target drug efficacy prediction model to process the physiological characteristic data based on the target drug efficacy prediction model. Then, the first drug efficacy prediction data of the target drug relative to the object to be tested can be obtained.

[0074] S280: Determine the drug applicable subjects corresponding to the target drug based on the first drug efficacy prediction data, the physiological characteristics and the target drug efficacy prediction model.

[0075] In this embodiment, when the first drug efficacy prediction data is obtained, the drug applicable subjects corresponding to the target drug can be determined according to the first drug efficacy data, the physiological characteristics and the target drug efficacy prediction model.

[0076] It should be noted that the target efficacy prediction model can be an interpretable model. It can also be an uninterpretable model. Among them, the interpretable model can be understood as a model whose decision-making process and prediction results can be understood and explained. In general, an interpretable model usually has a simple model structure, and its decision-making process is relatively intuitive and easy to understand. An uninterpretable model can be understood as a model whose decision-making process and the internal working principle of the model are difficult to understand and explain. This type of model usually has a complex structure and a large number of parameters, which makes its decision-making process highly abstract and nonlinear, making it difficult to be intuitively explained.

[0077] In this embodiment, when the target drug efficacy prediction model is an interpretable model, the drug efficacy impact parameters corresponding to the target drug under each physiological characteristic can be directly determined based on the first drug efficacy prediction data and the physiological characteristics. Furthermore, the drug applicable object corresponding to the target drug can be determined based on the drug efficacy impact parameters corresponding to each physiological characteristic. Exemplarily, at least one physiological characteristic whose drug efficacy impact parameter is higher than a preset drug efficacy impact threshold can be used as a screening condition for drug applicable objects.

[0078] In this embodiment, when the target drug efficacy prediction model is an uninterpretable model, a model interpretability method can be introduced to interpret the target drug efficacy prediction model. Furthermore, the drug applicable object of the target drug can be determined based on the interpretation result. Optionally, the model interpretability method can include feature importance analysis, local interpretability method or visualization method, etc.

[0079] As an optional implementation manner of the first embodiment of the present invention, when the model interpretability method adopted is feature importance analysis, based on the first pharmacodynamic prediction data, the physiological characteristic and the target pharmacodynamic prediction model, the drug applicable objects corresponding to the target drug are determined, including: for each physiological characteristic, the characteristic values ​​corresponding to the physiological characteristics other than the physiological characteristic in the physiological characteristic data are input into the target pharmacodynamic prediction model to obtain the second pharmacodynamic prediction data corresponding to the physiological characteristic; for each physiological characteristic, the pharmacodynamic influence parameter and the expected characteristic value corresponding to the physiological characteristic are determined according to multiple first pharmacodynamic prediction data and multiple second pharmacodynamic prediction data corresponding to the physiological characteristic; the target physiological characteristic is determined according to the multiple pharmacodynamic influence parameters, and the drug applicable objects corresponding to the target drug are determined based on the expected characteristic value corresponding to the target physiological characteristic.

[0080] Among them, the second pharmacodynamic prediction data can be obtained based on model prediction, and is used to characterize the pharmacodynamic situation characterized by the target drug on the object to be tested. It should be noted that the first pharmacodynamic prediction data is the pharmacodynamic data predicted by inputting all the acquired physiological characteristic data corresponding to the object to be tested into the target pharmacodynamic prediction model; the second pharmacodynamic prediction data is the pharmacodynamic data predicted by inputting the physiological characteristic data corresponding to the acquired object to be tested after removing a characteristic value corresponding to a physiological characteristic, and then inputting the removed physiological characteristic data into the target pharmacodynamic prediction model. It should be noted that the second pharmacodynamic prediction data corresponds to the physiological characteristic. In other words, for each object to be tested, the pharmacodynamic prediction data corresponding to the object to be tested may include the first pharmacodynamic prediction data and the second pharmacodynamic prediction data corresponding to each physiological characteristic.

[0081] In this embodiment, in order to analyze the degree of influence of each physiological characteristic in the physiological characteristic data on the efficacy of the target drug, it is possible to determine the efficacy prediction data obtained by processing the physiological characteristic data based on the target efficacy prediction model when the characteristic value corresponding to any physiological characteristic is removed from the physiological characteristic data.

[0082] In this embodiment, for each physiological characteristic, the characteristic value corresponding to the physiological characteristic can be eliminated from the physiological characteristic data corresponding to the object to be tested. Then, the physiological characteristic data after elimination can be input into the target drug efficacy prediction model to process the physiological characteristic data after elimination based on the target drug efficacy prediction model. Then, the second drug efficacy prediction data of the target drug relative to the object to be tested can be obtained, that is, the second drug efficacy prediction data corresponding to the physiological characteristic can be obtained.

[0083] Among them, the pharmacodynamic influence parameter is used to indicate the degree of influence of the physiological characteristics on the pharmacodynamics of the target drug. Among them, the pharmacodynamic influence parameter is used to indicate the degree of influence of the physiological characteristics on the pharmacodynamics of the target drug. The pharmacodynamic influence parameter and the degree of pharmacodynamic influence may be positively correlated. That is, the larger the pharmacodynamic influence parameter, the greater the degree of influence of the physiological characteristics on the pharmacodynamics of the target drug; the smaller the pharmacodynamic influence parameter, the smaller the degree of influence of the physiological characteristics on the pharmacodynamics of the target drug. The expected characteristic value can be understood as the characteristic value corresponding to the target drug when the target drug exhibits the best efficacy. In other words, when the characteristic value corresponding to the physiological characteristic is the expected characteristic value, the target drug can exhibit the best efficacy.

[0084] In this embodiment, for each object to be tested, when the first pharmacodynamic prediction data corresponding to the object to be tested and the second pharmacodynamic prediction data corresponding to each physiological characteristic of the object to be tested are obtained, the degree of pharmacodynamic influence of each physiological characteristic on the target drug can be analyzed according to the first pharmacodynamic prediction data and the second pharmacodynamic prediction data corresponding to each physiological characteristic. Further, for each physiological characteristic, the first pharmacodynamic prediction data corresponding to multiple objects to be tested and the second pharmacodynamic prediction data corresponding to multiple objects to be tested under the physiological characteristic can be processed. Then, the pharmacodynamic influence parameter and the expected characteristic value corresponding to the physiological characteristic can be obtained.

[0085] Optionally, based on multiple first pharmacodynamic prediction data and multiple second pharmacodynamic prediction data corresponding to the physiological characteristics, pharmacodynamic influence parameters and expected characteristic values ​​corresponding to the physiological characteristics are determined, including: processing the multiple first pharmacodynamic prediction data and the multiple second pharmacodynamic prediction data according to a predetermined characteristic evaluation algorithm to obtain the pharmacodynamic influence parameters and expected characteristic values ​​corresponding to the physiological characteristics.

[0086] The feature evaluation algorithm can be understood as an algorithm for evaluating the contribution of a certain feature to the final result. Optionally, the feature evaluation algorithm can be a Shapley value method.

[0087] As an optional implementation of the first embodiment of this invention, for each physiological characteristic, a plurality of first pharmacodynamic prediction data and a plurality of second pharmacodynamic prediction data corresponding to the physiological characteristic of the test object can be used as input parameters of the characteristic evaluation algorithm. Furthermore, the plurality of first pharmacodynamic prediction data and the plurality of second pharmacodynamic prediction data can be analyzed based on the characteristic evaluation algorithm to obtain the pharmacodynamic influence parameters and expected characteristic values ​​corresponding to the physiological characteristic.

[0088] In this embodiment, when the pharmacodynamic influence parameters corresponding to multiple physiological characteristics are obtained, the target physiological characteristic is determined according to the multiple pharmacodynamic influence parameters, and the drug applicable object corresponding to the target drug is determined based on the expected characteristic value corresponding to the target physiological characteristic.

[0089] The target physiological feature may be a physiological feature that meets the user's needs among the multiple physiological features included in the physiological feature data. The drug applicable object may be understood as the drug action object that enables the drug to exert the desired ideal drug effect.

[0090] In this embodiment, when the pharmacodynamic influence parameter and the expected characteristic value corresponding to each physiological characteristic are obtained, at least one target pharmacodynamic influence parameter can be screened out from multiple pharmacodynamic influence parameters according to preset screening criteria, and the physiological characteristic corresponding to the target pharmacodynamic influence parameter can be used as the target physiological characteristic.

[0091] Optionally, the target physiological characteristic is determined based on multiple pharmacodynamic influence parameter values, including: determining at least one pharmacodynamic influence parameter that exceeds a preset influence threshold from multiple pharmacodynamic influence parameters, and using the at least one pharmacodynamic influence parameter as the target pharmacodynamic influence parameter; determining the physiological characteristic corresponding to the target pharmacodynamic influence parameter, and using the physiological characteristic as the target physiological characteristic.

[0092] The preset impact threshold may be a pre-set parameter value used to limit the drug efficacy impact parameter. The preset impact threshold may be a user-defined setting or a default setting.

[0093] As an optional implementation of the first embodiment of this invention, each drug effect parameter can be compared with a preset effect threshold, and the drug effect parameters exceeding the preset effect threshold can be screened out, and the screened drug effect parameters can be used as target drug effect parameters. Furthermore, a physiological feature corresponding to the target drug effect parameter can be determined from multiple physiological features, and the physiological feature can be used as the target physiological feature.

[0094] In this embodiment, when the target physiological characteristic is determined, the drug-applicable subjects corresponding to the target drug can be screened according to the expected characteristic value corresponding to the target physiological characteristic, and when the characteristic value of the target physiological characteristic corresponding to any object is the corresponding expected characteristic value, the object can be regarded as the drug-applicable subject corresponding to the target drug.

[0095] As another optional implementation of this embodiment, when the model interpretability method adopted is a local interpretability method, the process of determining the drug applicable object corresponding to the target drug can be: arbitrarily select a target object to be tested from multiple objects to be tested, and simulate an approximate sample set based on the physiological characteristic data corresponding to the target object to be tested. Among them, the approximate sample set includes physiological characteristic data corresponding to multiple approximate objects to be tested. Afterwards, for each approximate object to be tested, the physiological characteristic data corresponding to the approximate object to be tested is input into the target drug efficacy prediction model to obtain the drug efficacy prediction data corresponding to the approximate object to be tested. Afterwards, the drug efficacy prediction model is trained based on the physiological characteristic data and drug efficacy prediction data corresponding to multiple approximate objects to be tested. Among them, the drug efficacy prediction model is a simple interpretable model. Afterwards, the model parameters of the drug efficacy prediction model can be used to explain the prediction of the target drug efficacy prediction model for each physiological characteristic corresponding to the target object to be tested. For example, the coefficient of drug efficacy prediction is analyzed. For example, if the coefficient of a physiological characteristic in the drug efficacy prediction model is positive and large, then the physiological characteristic may play an important role in the prediction of the target drug efficacy prediction model on the target test object. In this case, the physiological characteristic can be used as a screening condition for drug applicable objects. Otherwise, it may have a negative impact on the prediction. In this case, the physiological characteristic is removed from the screening condition for drug applicable objects.

[0096] The technical solution of the embodiment of the present invention is to obtain physiological characteristic data corresponding to multiple objects to be tested; wherein the physiological characteristic data includes characteristic values ​​corresponding to multiple physiological characteristics, and then, for each object to be tested, the physiological characteristic data corresponding to the object to be tested is processed according to a pre-trained target pharmacodynamic prediction model corresponding to the target drug, to obtain first pharmacodynamic prediction data of the target drug relative to the object to be tested, and then, for each physiological characteristic, the characteristic values ​​corresponding to other physiological characteristics in the physiological characteristic data except the physiological characteristic are input into the target pharmacodynamic prediction model to obtain second pharmacodynamic prediction data corresponding to the physiological characteristic, and further, based on the first pharmacodynamic prediction data, the physiological characteristics and the target pharmacodynamic prediction model, the drug applicable object corresponding to the target drug is determined, thereby achieving accurate simulation of pharmacodynamic data before conducting clinical trials, and then, accurately determining the effect of the ideal pharmacodynamic object, and then, achieving the effect of improving the success rate and efficiency of drug development on the basis of reducing drug development costs.

[0097] Embodiment 3

[0098] Figure 3 Schematic diagram of a device for constructing a drug efficacy prediction model provided by Embodiment 3 of the present invention. Figure 3As shown, the device includes: a data acquisition module 310, a model construction module 320, a data simulation module 330, a data prediction module 340 and a model training module 350.

[0099] The data acquisition module 310 is used to acquire physiological characteristic data corresponding to a plurality of sample objects and the real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each of the sample objects; the model construction module 320 is used to construct a pharmacodynamic parameter prediction model corresponding to the target drug according to the plurality of physiological characteristic data, the real pharmacokinetic data and the real pharmacodynamic data; the data simulation module 330 is used to perform data simulation on the plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects; the data prediction module 340 is used to calculate the physiological characteristic data corresponding to each of the simulated sample objects. The physiological characteristic data corresponding to the simulated sample object is input into the pharmacodynamic parameter prediction model to obtain the simulated pharmacodynamic data of the target drug relative to the simulated sample object; the model training module 350 is used to construct a plurality of training samples according to the plurality of physiological characteristic data, the real pharmacodynamic data, the physiological characteristic data corresponding to the plurality of simulated sample objects, and the simulated pharmacodynamic data of the target drug relative to the simulated sample object, and obtain a target pharmacodynamic prediction model corresponding to the target drug based on the plurality of training samples, so as to determine the drug applicable object corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested.

[0100] The technical solution of the embodiment of the present invention obtains physiological characteristic data corresponding to multiple sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each sample object, and then, according to the multiple physiological characteristic data, the real pharmacokinetic data and the real pharmacodynamic data, constructs a pharmacodynamic parameter prediction model corresponding to the target drug, and then, according to a predetermined data simulation algorithm, performs data simulation on the multiple physiological characteristic data to obtain physiological characteristic data corresponding to multiple simulated sample objects, and then, for each simulated sample object, inputs the physiological characteristic data corresponding to the simulated sample object into the pharmacodynamic parameter prediction model to obtain simulated pharmacodynamic data of the target drug relative to the simulated sample object, and then, according to the multiple simulated sample objects, Corresponding physiological characteristic data and simulated pharmacodynamic data of the target drug relative to the simulated sample object are used to construct multiple training samples, and a target pharmacodynamic prediction model corresponding to the target drug is obtained based on the multiple training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested. This solves the problem in the related technology that clinical trials fail due to lack of drug efficacy data or inaccurate selection of subjects, thereby affecting drug development efficiency and increasing drug development costs, and achieves accurate simulation of pharmacodynamic data before conducting clinical trials, thereby accurately determining the effect of ideal pharmacodynamic objects, thereby achieving the effect of improving drug development success rate and development efficiency on the basis of reducing drug development costs.

[0101] Optionally, the model building module 320 includes: a mapping relationship determination unit, a sub-model determination unit and a model building unit.

[0102] A mapping relationship determination unit, used to determine a mapping relationship between a plurality of the physiological characteristic data and a plurality of the real pharmacokinetic data, and use the mapping relationship as a pharmacokinetic parameter prediction submodel;

[0103] a sub-model determination unit, configured to determine at least one candidate pharmacokinetic-pharmacodynamic binding sub-model corresponding to the target drug according to the drug action mechanism corresponding to the target drug, and determine the pharmacokinetic-pharmacodynamic binding sub-model according to the real pharmacokinetic data, the real pharmacodynamic data and the at least one candidate pharmacokinetic-pharmacodynamic binding sub-model;

[0104] The model building unit is used to connect the pharmacokinetic parameter prediction sub-model and the pharmacokinetic-pharmacodynamic combination sub-model to obtain a pharmacodynamic parameter prediction model.

[0105] Optionally, the sub-model determination unit includes: a drug efficacy prediction data determination sub-unit, a drug efficacy data evaluation sub-unit and a sub-model determination sub-unit.

[0106] A drug efficacy prediction data determination subunit, for inputting the real pharmacokinetic data into the candidate drug efficacy combination submodel for each of the candidate drug efficacy combination submodels, to obtain drug efficacy prediction data corresponding to the target drug;

[0107] A drug efficacy data evaluation subunit, used to process the drug efficacy prediction data and the actual drug efficacy data corresponding to the target drug according to at least one preset evaluation index, and obtain the evaluation result corresponding to the candidate pharmacokinetic drug efficacy combination submodel under each of the preset evaluation indicators; wherein the preset evaluation indicator is used to evaluate the model prediction performance of the candidate pharmacokinetic drug efficacy combination submodel;

[0108] The sub-model determination sub-unit is used to determine a pharmacokinetic-pharmacodynamic binding sub-model from the at least one candidate pharmacokinetic-pharmacodynamic binding sub-model according to at least one of the evaluation results.

[0109] Optionally, the model training module 350 includes: a model processing unit and a model parameter correction unit.

[0110] A model processing unit, for inputting the physiological characteristic data corresponding to the simulated sample objects in the training samples into the drug efficacy prediction model to be trained, for a plurality of training samples, to obtain drug efficacy prediction data corresponding to the training samples;

[0111] A model parameter correction unit is used to determine a loss value based on the simulated efficacy data and the efficacy prediction data in the training sample, and to correct the model parameters in the efficacy prediction model to be trained based on the loss value until the loss function in the efficacy prediction model to be trained converges, thereby obtaining a target efficacy prediction model corresponding to the target drug.

[0112] Optionally, the device further includes: a physiological characteristic data acquisition module, a drug efficacy prediction module and a drug applicable object determination module.

[0113] A physiological characteristic data acquisition module, used to acquire physiological characteristic data corresponding to a plurality of objects to be tested; wherein the physiological characteristic data includes characteristic values ​​corresponding to a plurality of physiological characteristics;

[0114] A drug efficacy prediction module is used to process the physiological characteristic data corresponding to the test object according to a pre-trained target drug efficacy prediction model corresponding to the target drug for each test object, so as to obtain the first drug efficacy prediction data of the target drug relative to the test object; wherein the target drug efficacy prediction model is trained based on the physiological characteristic data corresponding to the sample object and the real drug efficacy data of the target drug relative to the sample object;

[0115] The drug applicable object determination module is used to determine the drug applicable object corresponding to the target drug based on the first drug efficacy prediction data, the physiological characteristics and the target drug efficacy prediction model.

[0116] Optionally, the target drug efficacy prediction model is an unexplainable model, and the drug applicable object determination module includes: a second drug efficacy prediction data determination unit, a drug efficacy influence parameter determination unit and a drug applicable object determination unit.

[0117] a second drug efficacy prediction data determination unit, for each of the physiological characteristics, inputting a characteristic value corresponding to other physiological characteristics in the physiological characteristic data except the physiological characteristic into the target drug efficacy prediction model to obtain second drug efficacy prediction data corresponding to the physiological characteristic;

[0118] a drug efficacy influence parameter determination unit, configured to determine, for each of the physiological characteristics, a drug efficacy influence parameter and an expected characteristic value corresponding to the physiological characteristic according to a plurality of the first drug efficacy prediction data and a plurality of the second drug efficacy prediction data corresponding to the physiological characteristic; wherein the drug efficacy influence parameter is used to indicate the degree of influence of the physiological characteristic on the drug efficacy of the target drug;

[0119] The drug applicable object determination unit is used to determine the target physiological characteristic according to the multiple drug efficacy influencing parameters, and determine the drug applicable object corresponding to the target drug based on the expected characteristic value corresponding to the target physiological characteristic.

[0120] Optionally, the pharmacodynamic effect parameter determination unit is specifically used to process the plurality of the first pharmacodynamic effect prediction data and the plurality of the second pharmacodynamic effect prediction data according to a predetermined feature evaluation algorithm to obtain the pharmacodynamic effect parameter and expected feature value corresponding to the physiological feature.

[0121] Optionally, the drug applicable object determination unit includes: a target parameter determination subunit and a drug applicable characteristic determination subunit.

[0122] a target parameter determination subunit, configured to determine at least one pharmacodynamic influence parameter exceeding a preset influence threshold value from the plurality of pharmacodynamic influence parameters, and use the at least one pharmacodynamic influence parameter as a target pharmacodynamic influence parameter;

[0123] The drug applicable characteristic determination subunit is used to determine the physiological characteristics corresponding to the target drug efficacy influencing parameters and use the physiological characteristics as target physiological characteristics.

[0124] The device for determining applicable drug objects provided in the embodiment of the present invention can execute the method for determining applicable drug objects provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0125] Embodiment 4

[0126] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0127] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0128] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0129] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, such as a method for determining a drug applicable subject.

[0130] In some embodiments, the method for determining a suitable drug object may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for determining a suitable drug object described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the method for determining a suitable drug object by any other appropriate means (e.g., by means of firmware).

[0131] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0132] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0133] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0134] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0135] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0136] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0137] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0138] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing a drug efficacy prediction model, characterized in that: include: Acquiring physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of a target drug relative to each of the sample objects; Constructing a drug efficacy parameter prediction model corresponding to the target drug according to the plurality of physiological characteristic data, the real pharmacokinetic data and the real drug efficacy data; Performing data simulation on the plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects; For each of the simulated sample objects, inputting the physiological characteristic data corresponding to the simulated sample object into the drug efficacy parameter prediction model to obtain the simulated drug efficacy data of the target drug relative to the simulated sample object; A plurality of training samples are constructed according to the physiological characteristic data corresponding to the plurality of simulated sample objects and the simulated pharmacodynamic data of the target drug relative to the simulated sample objects, and a target pharmacodynamic prediction model corresponding to the target drug is obtained based on the training of the plurality of training samples, so as to determine the drug applicable objects corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested.

2. The method for constructing a drug efficacy prediction model according to claim 1, characterized in that: The step of constructing a drug efficacy parameter prediction model corresponding to the target drug based on the plurality of physiological characteristic data, the real pharmacokinetic data and the real drug efficacy data comprises: Determining a mapping relationship between a plurality of the physiological characteristic data and a plurality of the real pharmacokinetic data, and using the mapping relationship as a pharmacokinetic parameter prediction submodel; Determining at least one candidate pharmacokinetic-pharmacodynamic binding submodel corresponding to the target drug according to the drug action mechanism corresponding to the target drug, and determining a pharmacokinetic-pharmacodynamic binding submodel according to the real pharmacokinetic data, the real pharmacodynamic data and the at least one candidate pharmacokinetic-pharmacodynamic binding submodel; The pharmacokinetic parameter prediction sub-model and the pharmacokinetic-pharmacodynamic binding sub-model are connected to obtain a pharmacodynamic parameter prediction model.

3. The method for constructing a drug efficacy prediction model according to claim 2, characterized in that: Determining the pharmacokinetic-pharmacodynamic binding submodel according to the real pharmacokinetic data, the real pharmacodynamic data and the at least one candidate pharmacokinetic-pharmacodynamic binding submodel comprises: For each of the candidate pharmacokinetic-pharmacodynamic combination submodels, the real pharmacokinetic data is input into the candidate pharmacokinetic-pharmacodynamic combination submodel to obtain the pharmacodynamic prediction data corresponding to the target drug; The pharmacodynamic prediction data and the actual pharmacodynamic data corresponding to the target drug are processed according to at least one preset evaluation index to obtain the evaluation result corresponding to the candidate pharmacodynamic combination submodel under each of the preset evaluation indicators; wherein the preset evaluation indicator is used to evaluate the model prediction performance of the candidate pharmacodynamic combination submodel; A pharmacodynamic binding submodel is determined from the at least one candidate pharmacodynamic binding submodel according to at least one of the evaluation results.

4. The method for constructing a drug efficacy prediction model according to claim 1, characterized in that: The target drug efficacy prediction model corresponding to the target drug is obtained by training based on the multiple training samples, including: For a plurality of training samples, inputting the physiological characteristic data corresponding to the simulated sample objects in the training samples into the drug efficacy prediction model to be trained to obtain drug efficacy prediction data corresponding to the training samples; According to the simulated pharmacodynamic data and the pharmacodynamic prediction data in the training sample, a loss value is determined, and the model parameters in the pharmacodynamic prediction model to be trained are corrected based on the loss value until the loss function in the pharmacodynamic prediction model to be trained converges, thereby obtaining a target pharmacodynamic prediction model corresponding to the target drug.

5. The method for constructing a drug efficacy prediction model according to claim 1, characterized in that: Also includes: Acquire physiological characteristic data corresponding to a plurality of objects to be tested; wherein the physiological characteristic data includes characteristic values ​​corresponding to a plurality of physiological characteristics; For each of the objects to be tested, the physiological characteristic data corresponding to the object to be tested is processed according to a pre-trained target pharmacodynamic prediction model corresponding to the target drug, so as to obtain first pharmacodynamic prediction data of the target drug relative to the object to be tested; wherein the target pharmacodynamic prediction model is trained based on the physiological characteristic data corresponding to the sample object and the real pharmacodynamic data of the target drug relative to the sample object; Based on the first drug efficacy prediction data, the physiological characteristics and the target drug efficacy prediction model, a drug applicable object corresponding to the target drug is determined.

6. The method for constructing a drug efficacy prediction model according to claim 5, characterized in that: The target drug efficacy prediction model is an unexplainable model, and determining the drug applicable object corresponding to the target drug based on the first drug efficacy prediction data, the physiological characteristics and the target drug efficacy prediction model includes: For each of the physiological characteristics, inputting characteristic values ​​corresponding to other physiological characteristics in the physiological characteristic data except the physiological characteristic into the target drug efficacy prediction model to obtain second drug efficacy prediction data corresponding to the physiological characteristic; For each of the physiological characteristics, a pharmacodynamic influence parameter and an expected characteristic value corresponding to the physiological characteristic are determined according to a plurality of the first pharmacodynamic prediction data and a plurality of the second pharmacodynamic prediction data corresponding to the physiological characteristic; wherein the pharmacodynamic influence parameter is used to indicate the degree of influence of the physiological characteristic on the pharmacodynamic of the target drug; A target physiological characteristic is determined according to a plurality of the pharmacodynamic influencing parameters, and a drug applicable object corresponding to the target drug is determined based on an expected characteristic value corresponding to the target physiological characteristic.

7. The method for constructing a drug efficacy prediction model according to claim 6, characterized in that: The determining, based on the plurality of the first drug effect prediction data and the plurality of the second drug effect prediction data corresponding to the physiological characteristics, the drug effect influence parameter and the expected characteristic value corresponding to the physiological characteristics comprises: The plurality of first drug efficacy prediction data and the plurality of second drug efficacy prediction data are processed according to a predetermined characteristic evaluation algorithm to obtain drug efficacy influence parameters and expected characteristic values ​​corresponding to the physiological characteristics.

8. The method for constructing a drug efficacy prediction model according to claim 6, characterized in that: Determining the target physiological characteristic according to the plurality of pharmacodynamic effect parameter values ​​comprises: Determine at least one pharmacodynamic influence parameter exceeding a preset influence threshold value from the plurality of pharmacodynamic influence parameters, and use the at least one pharmacodynamic influence parameter as a target pharmacodynamic influence parameter; The physiological characteristics corresponding to the target drug efficacy influence parameters are determined, and the physiological characteristics are used as target physiological characteristics.

9. A drug efficacy prediction model construction device, characterized in that: include: A data acquisition module, used to acquire physiological characteristic data corresponding to a plurality of sample objects and real pharmacokinetic data and real pharmacodynamic data of the target drug relative to each of the sample objects; A model building module, used to build a pharmacodynamic parameter prediction model corresponding to the target drug according to the plurality of physiological characteristic data, the real pharmacokinetic data and the real pharmacodynamic data; A data simulation module, used to perform data simulation on the plurality of physiological characteristic data according to a predetermined data simulation algorithm to obtain physiological characteristic data corresponding to a plurality of simulated sample objects; A data prediction module, for inputting the physiological characteristic data corresponding to each simulated sample object into the drug efficacy parameter prediction model to obtain the simulated drug efficacy data of the target drug relative to the simulated sample object; A model training module is used to construct a plurality of training samples based on the plurality of physiological characteristic data, the real pharmacodynamic data, the physiological characteristic data corresponding to the plurality of simulated sample objects, and the simulated pharmacodynamic data of the target drug relative to the simulated sample objects, and to obtain a target pharmacodynamic prediction model corresponding to the target drug based on the plurality of training samples, so as to determine a drug applicable object corresponding to the target drug based on the target pharmacodynamic prediction model and the physiological characteristic data corresponding to the object to be tested.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for constructing a drug efficacy prediction model according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for constructing a drug efficacy prediction model according to any one of claims 1 to 8 when executed.

Images