Drug efficacy and action mechanism prediction method and system based on Raman spectrum

Abstract

The invention discloses a drug efficacy and action mechanism prediction method and system based on Raman spectrum, and relates to the technical field of drug efficacy prediction.The method comprises the following steps that a to-be-detected sample is prepared, Raman spectrum scanning is conducted on the to-be-detected sample, and corresponding Raman spectrum data is obtained; processing the Raman spectrum data to form a spectrum data set, and dividing the spectrum data set to obtain a corresponding training set and a test set; constructing a prediction model, training the prediction model by using the training set, and performing prediction effect evaluation on the prediction model by using the test set; acquiring sample data of other drugs, and inputting the sample data into the prediction model for prediction; according to the method, cells or cell secretions stimulated by different types of immune drugs are analyzed through a machine learning algorithm, Raman spectrum databases and models of different types of drugs are established, and powerful support is provided for research and development of immune drugs and further mechanism exploration.

Description

Technical Field

[0001] The present invention relates to the technical field of drug efficacy prediction, and more particularly to a method and system for predicting drug efficacy and action mechanism based on Raman spectroscopy. Background Art

[0002] Currently, immune cells play a key role in maintaining the body's health, and regulating immune cell status is considered a potential therapeutic strategy for a variety of diseases, such as certain infections, tumors, and autoimmune diseases. Regulating immune cell function through drugs has important clinical significance.

[0003] The research on the pathways / mechanisms of drug action on immune cells currently mainly includes RNA sequencing, proteomics, flow cytometry and other research methods.

[0004] (1) RNA sequencing is based on high-throughput sequencing technology. It detects changes in gene expression in cells under the action of drugs, namely the RNA transcriptome, and quantifies the expression level of genes to reveal the mechanism of action of drugs. This method has the disadvantages of complex data processing, high professionalism, reliance on bioinformatics analysis, and batch differences. In addition, transcription information does not equal function, and processes such as non-coding RNA and post-transcriptional modification are ignored. In addition, it is expensive and requires professional equipment and technical support.

[0005] (2) Proteomics has also been used for drug efficacy analysis. For example, liquid chromatography-mass spectrometry (LC / MS-MS) separates protein samples by liquid chromatography and uses mass spectrometry to analyze different peptides to infer the types and quantities of proteins in the sample. The main problems with this method are that the proteome is too complex, protein identification relies on databases, low-abundance proteins are ignored, and post-translational modifications of proteins are difficult to identify.

[0006] (3) Flow cytometry uses fluorescently labeled molecular markers on the cell surface or within the cell. By analyzing the flow cytometry results, it can detect various cell states such as cell proliferation, cell death, phenotypic changes, and signal pathway activation after drug administration to reveal the efficacy of the drug. The main problems with this method are its reliance on labeled antibodies and the large heterogeneity of fluorescent channel types and batches.

[0007] Therefore, how to provide a drug efficacy prediction method that can solve the above problems is an urgent problem that those skilled in the art need to solve. Summary of the Invention

[0008] In view of this, the present invention provides a method and system for predicting drug efficacy and mechanism of action based on Raman spectroscopy. A Raman spectrometer is used to measure the Raman spectra of cells or cell secretions. The cells or cell secretions after stimulation by different types of immune drugs are analyzed by a machine learning algorithm to establish a Raman spectrum database and model for different types of drugs. Using this database and model, the Raman spectra of cells and cell secretions after stimulation by other drugs are analyzed and predicted, thereby predicting the effect of the drug on immune cells and its mechanism of action. The purpose of this method is to obtain the drug efficacy and mechanism of action of immune drugs on immune cells more quickly and accurately, and also to provide strong support for the research and development of immune drugs and further mechanism exploration.

[0009] In order to achieve the above object, the present invention adopts the following technical solutions:

[0010] A method for predicting drug efficacy and mechanism of action based on Raman spectroscopy, comprising the following steps:

[0011] S1: Prepare a sample to be tested, and perform Raman spectrum scanning on the sample to be tested to obtain corresponding Raman spectrum data;

[0012] S2: Process the Raman spectral data to form a spectral dataset, and divide the spectral dataset into corresponding training sets and test sets;

[0013] S3: Build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set;

[0014] S4: Obtain sample data of other drugs, and input the sample data into the prediction model obtained in S3 for prediction.

[0015] Preferably, it also includes:

[0016] S5: Using the trained prediction model to predict the sample data, and obtain the corresponding drug action mechanism.

[0017] Preferably, the S1 specifically includes:

[0018] S11: Preparation of dry probe;

[0019] S12: Dropping the cell secretions after drug stimulation onto the dry probe, and then collecting Raman spectrum information of the cell secretions corresponding to different drugs.

[0020] Preferably, the S2 specifically includes:

[0021] S21: preprocessing Raman spectroscopy data;

[0022] S22: Divide the Raman spectrum data obtained in S21 into a training set and a test set according to a preset ratio.

[0023] Preferably, the S3 specifically includes:

[0024] S31: constructing a prediction model, wherein the prediction model is any one of an SVM model and a PLS-DA processing model;

[0025] S32: Label the training set into categories and put the training set into the prediction model for training;

[0026] S33: Put the test set into the prediction model for model evaluation.

[0027] The present invention also provides a drug efficacy and mechanism prediction system based on Raman spectroscopy, comprising:

[0028] The spectrum acquisition module is used to prepare the sample to be tested and perform Raman spectrum scanning on the sample to be tested to obtain the corresponding Raman spectrum data;

[0029] The data preprocessing module is used to process the Raman spectral data to form a spectral data set, and divide the spectral data set into corresponding training sets and test sets;

[0030] The model building module is used to build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set;

[0031] The prediction module is used to obtain sample data of other drugs and input the sample data into the prediction model for prediction.

[0032] Preferably, the spectrum acquisition module includes: a sample preparation unit and a spectrum acquisition unit.

[0033] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for predicting drug efficacy and mechanism of action based on Raman spectroscopy, which has the following beneficial effects:

[0034] 1. High sensitivity and high specificity: Compared with traditional methods, the method provided by the present invention can efficiently capture subtle phenotypic changes of immune cells under drug stimulation, provide higher sensitivity and specificity, and avoid the damage of traditional detection methods during cell activity and sample processing.

[0035] 2. Non-destructive testing: The present invention uses Raman spectroscopy technology for non-destructive testing, avoiding the disadvantage of traditional methods that require destroying cell structure or activity, and can monitor the effects of drugs on immune cells in real time and continuously.

[0036] 3. Comprehensive mechanism of action analysis: This invention uses machine learning algorithms to perform multi-dimensional analysis of Raman spectroscopy data, which can not only predict the immunomodulatory effects of drugs but also reveal their potential mechanisms of action, especially the effects of drugs on the intracellular and extracellular biological processes of immune cells.

[0037] 4. Fast and efficient drug screening and optimization: Based on the combination of Raman spectroscopy and machine learning, this invention can efficiently screen a large number of drugs, helping drug developers to quickly identify potential drugs and optimize dosages, significantly improving the efficiency of drug development.

[0038] 5. Can be widely used in immunology research and drug development: The prediction model provided by this invention is not only suitable for the development of immunomodulatory drugs, but can also be applied in other immunology research fields, providing important support for immunotherapy, vaccine development, etc.

[0039] In summary, the present invention combines Raman spectroscopy with machine learning to provide an efficient and accurate drug response prediction tool, which can achieve significant technical, economic, and social benefits in drug development, immune mechanism exploration, disease treatment, and other aspects, providing important support for innovation and development in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0041] Figure 1 This is an overall flow chart of a method for predicting drug efficacy and mechanism of action based on Raman spectroscopy provided by the present invention;

[0042] Figure 2 Average Raman spectra of representative drugs with different efficacy / mechanisms provided in embodiments of the present invention;

[0043] Figure 3A Schematic diagram of the classification results of the test set and training set of the prediction model (PLS-DA, SVM) provided in the embodiment of the present invention for drug efficacy prediction;

[0044] Figure 3B Schematic diagram of the efficacy prediction results of the prediction set CU-T12-9 and the training set for drug efficacy prediction using the prediction model (PLS-DA, SVM) provided in an embodiment of the present invention;

[0045] Figure 3CSchematic diagram of the classification results of the test set and training set of the prediction model (PLS-DA, SVM) provided in the embodiment of the present invention for drug action mechanism prediction;

[0046] Figure 3D Schematic diagram of the mechanism prediction results of the prediction set CU-T12-9 and the training set using the prediction model (PLS-DA, SVM) provided in an embodiment of the present invention for drug mechanism prediction;

[0047] Figure 4 This is a block diagram of the structural principle of a drug efficacy and action mechanism prediction system based on Raman spectroscopy provided by the present invention;

[0048] Figure 5 A schematic diagram of a spectrum acquisition unit provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0050] See also Figure 1 As shown, the embodiment of the present invention discloses a method for predicting drug efficacy and mechanism of action based on Raman spectroscopy, comprising the following steps:

[0051] S1: Prepare a sample to be tested, and perform Raman spectrum scanning on the sample to be tested to obtain corresponding Raman spectrum data;

[0052] S2: Process the Raman spectral data to form a spectral dataset, and divide the spectral dataset into corresponding training sets and test sets;

[0053] S3: Build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set;

[0054] S4: Obtain sample data of other drugs, and input the sample data into the prediction model obtained in S3 for prediction.

[0055] In a specific embodiment, it also includes:

[0056] S5: Using the trained prediction model to predict the sample data, and obtain the corresponding drug action mechanism.

[0057] In a specific embodiment, the S1 specifically includes:

[0058] S11: Preparation of dry probe;

[0059] S12: Dropping the cell secretions after drug stimulation onto the dry probe, and then collecting Raman spectrum information of the cell secretions corresponding to different drugs.

[0060] Specifically, S11 specifically includes: preparing a gold nanosphere solution; dropping the gold nanosphere solution onto a hard substrate, and then dropping the gold nanosphere solution again after drying, and repeating the process multiple times to obtain a dry probe.

[0061] Using a chloroauric acid redox method, microscopically spherical gold nanospheres with a size distribution of 20-100 nm were prepared. A corresponding solution was formed, and the absorbance of the prepared gold nanosphere liquid was measured at 520 nm, with an absorbance range of 0.1-10 OD. The rigid substrate could be any of silicon wafers, quartz wafers, aluminum foil, or glass. After drying, the gold nanosphere solution was added dropwise again, and this process was repeated 2-10 times.

[0062] Immune cells in EC of drugs with different efficacy / mechanisms 50 or IC 50 The supernatant secreted after 12 hours of stimulation at the concentration was placed on a SERS substrate for spectral collection, and the integration time of the Raman spectrum scan was 7 seconds.

[0063] In a specific embodiment, S2 specifically includes:

[0064] S21: preprocessing Raman spectroscopy data;

[0065] S22: Divide the Raman spectrum data obtained in S21 into a training set and a test set according to a preset ratio, where the preset ratio may be 2:1.

[0066] Specifically, the pre-processing process in S21 may include smoothing and baseline correction processing of the Raman spectrum data, wherein the smoothing process adopts SG smoothing and the baseline correction adopts airPLS.

[0067] In a specific embodiment, S3 specifically includes:

[0068] S31: constructing a prediction model, wherein the prediction model is any one of an SVM model and a PLS-DA processing model;

[0069] S32: Label the training set into categories and put the training set into the prediction model for training;

[0070] S33: Put the test set into the prediction model for model evaluation.

[0071] In a specific embodiment, a drug efficacy prediction model was developed using five drugs from three efficacy categories, including agonists (LPS and CPG), inhibitors (D95 and Sco), and placebo. A total of 1,350 spectral data sets were used to construct sample data for training. PLS-DA employed a ten-fold crossover operation for model parameter optimization, and SVM, in fitcecoc, used a support vector machine as the base classifier to train the model.

[0072] See also Figure 4 As shown, an embodiment of the present invention further provides a system using the method for predicting drug efficacy and mechanism of action based on Raman spectroscopy as described in any of the above embodiments, comprising:

[0073] The spectrum acquisition module is used to prepare the sample to be tested and perform Raman spectrum scanning on the sample to be tested to obtain the corresponding Raman spectrum data;

[0074] The data preprocessing module is used to process the Raman spectral data to form a spectral data set, and divide the spectral data set into corresponding training sets and test sets;

[0075] The model building module is used to build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set;

[0076] The prediction module is used to obtain sample data of other drugs and input the sample data into the prediction model for prediction.

[0077] In a specific embodiment, the spectrum acquisition module includes: a sample preparation unit and a spectrum acquisition unit.

[0078] For details, see Figure 5 As shown, the spectrum acquisition unit may include a laser 1 and a spectrometer 2. The laser 1 is used to emit laser light to the immune cell supernatant sample to be tested. The spectrometer 2 is used to collect Raman light scattered by the immune cell supernatant sample to obtain Raman spectrum data.

[0079] The data preprocessing module, model building module and prediction module can be implemented by computer 3. Computer 3 is connected to spectrometer 2 and is used to preprocess Raman spectral data, obtain preprocessed Raman spectral data, evaluate the drug efficacy / mechanism prediction model and predict unknown drug data; display 4 is connected to computer 3 and is used to display the prediction results of the drug efficacy / mechanism prediction model.

[0080] Furthermore, the apparatus may further include a fiber optic probe, which is connected to the laser 1 and the spectrometer 2, respectively, and is configured to transmit the laser light emitted by the laser 1 to the surface of the supernatant of the immune cells to be tested, and transmit the Raman light scattered by the surface of the supernatant of the immune cells to be tested to the spectrometer 2.

[0081] Furthermore, the device further comprises a movable slide, which is used to place the supernatant of the immune cells to be tested and to adjust the position and height of the supernatant of the immune cells to be tested.

[0082] In order to verify the effectiveness of the method provided by the embodiment of the present invention, the specific algorithm steps are as follows:

[0083] 1. Prediction of drug efficacy

[0084] (1) Data preparation: 900 spectral data of 5 drugs were spectrally preprocessed and divided into training set and test set with a quantity ratio of 2:1. The training set data were labeled as 1, 2, and 3 according to the efficacy type.

[0085] (2) Model optimization and training: Cross-fold operation (taking PLS-DA as an example): The training set data is randomly divided into 10 subsets, 9 of which are used as training sets each time, and the remaining 1 subset is used as the validation set, and the cycle is repeated 10 times.

[0086] Tuning hyperparameters: Optimize the model by selecting the hyperparameters that give the best performance during cross-validation.

[0087] Model generation: A drug efficacy prediction model is constructed based on the optimized parameters.

[0088] The drug efficacy prediction model constructed according to the above method has excellent prediction performance in the test set and the efficacy prediction of CU-T12-9 that did not participate in the model training. The average Raman spectra of agonists, inhibitors and placebos are shown in Figure 2. Figure 2 As shown in the figure, the accuracy of the efficacy prediction model is as follows Figure 3A -B.

[0089] Specifically, the drug efficacy prediction models constructed by PLS-DA and SVM respectively achieved an accuracy rate of more than 97% in the prediction of test set data; the drug efficacy prediction models constructed by PLS-DA and SVM respectively achieved an accuracy rate of more than 95% in the prediction of agonist CU-T12-9.

[0090] 2. Predictions on drug mechanisms

[0091] (1) Data preparation: Spectral preprocessing was performed on 360 spectral data of two drugs. The preprocessed data were divided into training set and test set with a quantity ratio of 2:1. The training set data were labeled as 1 and 2 according to the mechanism type.

[0092] (2) Model optimization and training: Ten-fold crossover operation (taking PLS-DA as an example): the training set data is randomly divided into 10 subsets, 9 of which are used as training sets each time, and the remaining 1 subset is used as the validation set, and the cycle is repeated 10 times.

[0093] Tuning hyperparameters: Optimize the model by selecting the hyperparameters that give the best performance during cross-validation.

[0094] Model generation: A drug mechanism prediction model is constructed based on the optimized parameters.

[0095] The drug mechanism prediction model constructed according to the above method has excellent prediction performance in the test set and the mechanism prediction of TLR1 / 2 agonist CU-T12-9 that did not participate in model training. The average Raman spectra of TLR1 / 2 agonist Pam3Csk4 and TLR4 agonist LPS are shown in Figure 2. Figure 2 As shown in Figure 2, the accuracy of the mechanism prediction model is as follows: Figure 3C -D as shown.

[0096] Specifically, the drug mechanism prediction models constructed by PLS-DA and SVM respectively achieved an accuracy of more than 99% in the prediction of test set data; the drug mechanism prediction models constructed by PLS-DA and SVM respectively achieved an accuracy of more than 96% in the prediction of CU-T12-9.

[0097] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0098] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting drug efficacy and mechanism of action based on Raman spectroscopy, characterized in that: The following steps are involved: S1: Prepare a sample to be tested, and perform Raman spectrum scanning on the sample to be tested to obtain corresponding Raman spectrum data; S2: Process the Raman spectral data to form a spectral dataset, and divide the spectral dataset into corresponding training sets and test sets; S3: Build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set; S4: Obtain sample data of other drugs, and input the sample data into the prediction model obtained in S3 for prediction.

2. The method for predicting drug efficacy and mechanism of action based on Raman spectroscopy according to claim 1, characterized in that: Also includes: S5: Using the trained prediction model to predict the sample data, and obtain the corresponding drug action mechanism.

3. The method for predicting drug efficacy and mechanism of action based on Raman spectroscopy according to claim 1, characterized in that Said S1 specifically includes: S11: Preparation of dry probe; S12: Dropping the cell secretions after drug stimulation onto the dry probe, and then collecting Raman spectrum information of the cell secretions corresponding to different drugs.

4. The method for predicting drug efficacy and mechanism of action based on Raman spectroscopy according to claim 1, wherein: The S2 specifically includes: S21: preprocessing Raman spectroscopy data; S22: Divide the Raman spectrum data obtained in S21 into a training set and a test set according to a preset ratio.

5. The method for predicting drug efficacy and mechanism of action based on Raman spectroscopy according to claim 1, wherein: The S3 specifically includes: S31: constructing a prediction model, wherein the prediction model is any one of an SVM model and a PLS-DA processing model; S32: Label the training set into categories and put the training set into the prediction model for training; S33: Put the test set into the prediction model for model evaluation.

6. A system using the method for predicting drug efficacy and mechanism of action based on Raman spectroscopy according to claims 1-5, characterized in that: include: The spectrum acquisition module is used to prepare the sample to be tested and perform Raman spectrum scanning on the sample to be tested to obtain the corresponding Raman spectrum data; The data preprocessing module is used to process the Raman spectral data to form a spectral data set, and divide the spectral data set into corresponding training sets and test sets; The model building module is used to build a prediction model, train the prediction model using the training set, and evaluate the prediction effect of the prediction model using the test set; The prediction module is used to obtain sample data of other drugs and input the sample data into the prediction model for prediction.

7. The system according to claim 6, characterized in that The spectrum acquisition module includes: a sample preparation unit and a spectrum acquisition unit.

Images