A biological radiation dose assessment method, device and terminal equipment

Abstract

This invention discloses a method, apparatus, and terminal device for assessing biological radiation dose. The method includes: extracting the expression information of a first target gene from a biological sample to be assessed; inputting the expression information of the first target gene into a preset dose assessment model, so that the model determines the radiation dose of the biological sample to be assessed based on the expression information of the first target gene; wherein, the model finds differentially expressed genes by comparing the transcriptome sequencing information of irradiated and corresponding unirradiated training samples, and performs feature correlation screening and quantitative gene amplification fluorescence detection on the differentially expressed genes to obtain the expression information of a second target gene; using the actual radiation dose of the irradiated training sample and the corresponding expression information of the second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain a trained dose assessment model. By implementing this invention, the radiation dose of the sample to be assessed can be predicted rapidly.

Description

Technical Field

[0001] This invention relates to the field of biological dose assessment, and more particularly to a biological radiation dose assessment method, apparatus, and terminal equipment. Background Technology

[0002] Currently, there are relatively few methods available for calculating the radiation dose received by an organism. Conventional methods include using genetic approaches, such as chromosomal aberrations, cytokinesis-block micronucleus, and fluorescence in situ hybridization, as well as various omics analyses, such as genomics, metabolomics, and proteomics, and identifying biomarkers for dose reconstruction.

[0003] While conventional radiation dose assessment methods provide important tools for scientific research, limitations in efficiency and cost-effectiveness pose challenges to their practical application. Conventional methods, such as genetic and omics analyses, require complex experimental procedures and data analysis, are time-consuming, and demand significant human and time investment. In emergency situations like radiation accidents, rapid and accurate dose assessment is crucial, but conventional methods often fail to meet the requirements of real-time performance and speed. Summary of the Invention

[0004] This invention provides a biological radiation dose assessment method, apparatus, and terminal device that can quickly predict the radiation dose of a sample to be assessed.

[0005] An embodiment of the present invention provides a method for assessing biological radiation dose, comprising:

[0006] Extract the expression information of the first target gene from the biological sample to be evaluated;

[0007] The expression information of the first target gene is input into a preset dose assessment model so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene.

[0008] The construction of the dose assessment model includes:

[0009] Obtain several sets of training samples of subjects subjected to known radiation doses and training samples of subjects not subjected to radiation.

[0010] Transcriptome sequencing was performed on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples.

[0011] The transcriptome sequencing information of the irradiated training samples was compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes.

[0012] By screening for the correlation of features of differentially expressed genes, a second target gene is obtained;

[0013] Quantitative gene amplification fluorescence detection was performed on the second target gene to obtain its expression information;

[0014] Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model.

[0015] Furthermore, correlation screening of differentially expressed genes is performed to obtain second target genes, including:

[0016] For each differentially expressed gene, calculate the logarithmic fold change, significance level, and correlation coefficient.

[0017] Differentially expressed genes that simultaneously meet the second, third, and fourth preset conditions are selected from the differentially expressed genes and used as the second target genes for preliminary screening. The second preset condition is: the absolute value of the logarithmic fold change is greater than the first preset threshold; the third preset condition is: the significance level is less than the second preset threshold; and the fourth preset condition is: the absolute value of the correlation coefficient is greater than or equal to the third preset threshold.

[0018] The Boruta algorithm was used to perform feature screening on the initially screened second target genes to obtain the final second target genes.

[0019] Furthermore, using the actual radiation dose of the irradiated training samples and the corresponding expression information of the second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model, including:

[0020] The irradiated training samples are divided into a first training set and a first test set according to a preset ratio;

[0021] The actual radiation dose of each training sample in the first training set and the expression information of the corresponding second target gene are input into the dose assessment model to be trained for training until the number of training times exceeds the preset number, and the trained dose assessment model is obtained.

[0022] Furthermore, the dose assessment model includes: a basic sub-model and an optimized sub-model;

[0023] The actual radiation dose of each training sample in the first training set and the corresponding expression information of the second target gene are input into the dose assessment model to be trained for training until the first training iteration exceeds the first preset iteration, resulting in a trained dose assessment model, including:

[0024] Divide the first training set into several training subsets;

[0025] During each training session, one training subset is selected sequentially from all training subsets as the second test set, and the remaining training subsets are used as the second training set; the training subsets can be selected cyclically.

[0026] The basic sub-model is trained based on the data in the second training set until the second training number exceeds the second preset number, so as to obtain the trained basic sub-model and the first predicted radiation dose obtained by each training of the basic sub-model.

[0027] Based on the first predicted radiation doses obtained from each training of the basic sub-model and the corresponding data of each first predicted radiation dose in the second training set, the optimized sub-model is trained until the third training number exceeds the third preset number, and the trained optimized sub-model is obtained.

[0028] The second test set is input into the trained base sub-model and the trained optimized sub-model to obtain the predicted radiation dose for each data in the second test set.

[0029] Based on the predicted radiation dose of each data in the second test set and the actual radiation dose corresponding to each data in the second test set, the values ​​of the coefficient of determination and root mean square error are calculated.

[0030] If the number of training iterations is less than the first preset number of iterations, then continue with the next training iteration; otherwise, compare the values ​​of the coefficient of determination and root mean square error obtained from each training iteration, and take the dose assessment model with the largest coefficient of determination and the smallest root mean square error as the completed dose assessment model.

[0031] Furthermore, the base sub-models include combinations of the following regression sub-models: Random Forest Regression Sub-model, Linear Regression Sub-model, XGBoost Regression Sub-model, CatBoost Regression Sub-model, Elastic Network Regression Sub-model, and Support Vector Regression Sub-model.

[0032] Furthermore, the basic sub-model is trained based on the data in the second training set until the second training iterations exceed the second preset number of iterations, resulting in the trained basic sub-model and the first predicted radiation doses obtained from each training iteration of the basic sub-model, including:

[0033] The basic sub-model is trained based on the data in the second training set; wherein, during each training, for each regression sub-model, the expression information of the second target gene of each data in the second training set is used as input and the first predicted radiation dose of each data in the second training set is used as output for training.

[0034] For each regression sub-model, the root mean square error of each regression sub-model is calculated based on the first predicted radiation dose and the corresponding actual radiation dose for each data in each second training set. The root mean square errors of each regression sub-model are then weighted to obtain the first loss function of the base sub-model.

[0035] Determine if the second number of training iterations is less than the second preset number of iterations. If so, repeat the next training iteration. If not, compare the first loss function values ​​obtained from each training iteration and use the base sub-model with the smallest first loss function value as the completed base sub-model.

[0036] Furthermore, based on the first predicted radiation doses obtained from each training iteration of the base sub-model and the corresponding data in the second training set, the optimized sub-model is trained until the third training iteration exceeds the third preset iteration, resulting in a fully trained optimized sub-model, including:

[0037] Each first predicted radiation dose obtained from each training of the basic sub-model is divided into several groups of predicted data; among them, the predicted data includes: each first predicted radiation dose obtained from the same training of the basic sub-model.

[0038] The optimized sub-model is trained based on the first predicted radiation doses obtained from each training of the base sub-model and the corresponding data of each first predicted radiation dose in the second training set. In each training session, a set of predicted data is selected sequentially from several sets of predicted data. For each first predicted radiation dose in the set of predicted data and the corresponding data of the first predicted radiation dose in the second training set, the training is performed with the first predicted radiation dose and the corresponding data of the first predicted radiation dose in the second training set as inputs and the second predicted radiation dose of the second training set data as outputs.

[0039] The second loss function is calculated based on the second predicted radiation dose and the corresponding actual radiation dose for each second training set data.

[0040] Determine if the third training iteration is less than the third preset iteration. If so, continue with the next training iteration. If not, compare the second loss function values ​​obtained from each training iteration and select the optimized sub-model with the smallest second loss function value as the completed optimized sub-model.

[0041] Furthermore, based on the predicted radiation dose of each data point in the second test set and the corresponding actual radiation dose, the values ​​of the coefficient of determination and root mean square error are calculated, including:

[0042] Based on the predicted radiation dose and the corresponding actual radiation dose for each data point in the second test set, the coefficient of determination is calculated using the following formula:

[0043]

[0044] Based on the predicted radiation dose and the corresponding actual radiation dose of each data point in the second test set, the root mean square error is calculated using the following formula:

[0045]

[0046] Among them, R 2 Y represents the coefficient of determination, RMSE represents the root mean square error, and Y represents the root mean square error. i This represents the actual radiation dose of the i-th data point in the second test set. This represents the predicted radiation dose for the i-th data point in the second test set. denoted as the average actual radiation dose in the second test set, and n represents the total amount of data in the test set.

[0047] Based on the above method embodiments, the present invention provides corresponding device embodiments, including: a first target gene expression information extraction module and a dose assessment module;

[0048] The first target gene expression information extraction module is used to extract the expression information of the first target gene in the biological sample to be evaluated;

[0049] The dose assessment module is used to input the expression information of the first target gene into a preset dose assessment model so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene.

[0050] The construction of the dose assessment model includes:

[0051] Obtain several sets of training samples of irradiated and unirradiated subjects;

[0052] Transcriptome sequencing was performed on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples.

[0053] The transcriptome sequencing information of the irradiated training samples was compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes.

[0054] By screening for the correlation of features of differentially expressed genes, a second target gene is obtained;

[0055] Quantitative gene amplification fluorescence detection was performed on the second target gene to obtain its expression information;

[0056] Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model.

[0057] Based on the above method embodiments, the present invention provides a corresponding terminal device embodiment, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the steps of the biological radiation dose assessment method as described in the present invention.

[0058] Compared with the prior art, the beneficial effects of this embodiment are as follows:

[0059] This invention obtains the expression information of specific target genes in the biological sample to be evaluated. Compared with traditional gene detection methods, the target genes of this invention are pre-determined to be genes associated with radiation dose, eliminating the step of identifying differentially expressed genes caused by radiation by comparing with normal biological samples, thus improving the efficiency of dose assessment. Then, the expression information of the specific target genes is input into a preset dose assessment model, so that the dose assessment model determines the radiation dose of the sample to be evaluated based on the expression information of the specific target genes. Through a large amount of sample data, the model can learn the relationship between different gene expressions and their corresponding radiation doses, thereby enabling radiation dose assessment in a shorter time. This invention can quickly predict the radiation dose of the sample to be evaluated based on the relationship between differentially expressed genes and radiation doses already learned by the dose assessment model. Attached Figure Description

[0060] Figure 1 This is a schematic flowchart of a biological radiation dose assessment method provided in an embodiment of the present invention;

[0061] Figure 2 This is a schematic flowchart of the training process of a dose assessment model provided in an embodiment of the present invention;

[0062] Figure 3 This is a dose reconstruction diagram of the dose assessment model based on the first test set provided in an embodiment of the present invention;

[0063] Figure 4 This is a schematic diagram of the structure of a biological radiation dose assessment device provided in an embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated.

[0066] like Figure 1 As shown, an embodiment of the present invention provides a method for assessing biological radiation dose, which includes at least the following steps:

[0067] Step S1: Extract the expression information of the first target gene from the biological sample to be evaluated;

[0068] For step S1, it is first necessary to obtain the biological sample of the subject to be evaluated. The biological sample can be the subject's blood, urine, hair, oral swab, or other biological samples that can provide genetic information about the subject. In this embodiment, mouse blood samples are used as an example to obtain blood from mice exposed to unknown radiation.

[0069] Then, the expression information of the first target gene in the blood of mice exposed to unknown radiation was measured using quantitative real-time PCR (qPCR). qPCR is a highly sensitive and quantitative method that can accurately measure gene expression levels, and the results obtained have high precision and reproducibility. In this embodiment, primers and probes were designed according to the sequence of the first target gene, RNA was extracted and reverse transcribed to synthesize cDNA, and then amplification was performed using a qPCR instrument. By monitoring the fluorescence signal, the progress and results of the PCR products were tracked, and the expression level of the first target gene was calculated accordingly.

[0070] It should be noted that the first target gene involved in this invention is a gene that has been pre-determined to be associated with radiation dose. Therefore, in actual biological radiation dose assessment, these radiation dose-associated genes can be directly extracted, and the expression levels of these target genes can be calculated using qPCR technology. Compared with traditional gene detection methods, this invention eliminates the need to identify differentially affected genes by radiation through comparison with normal biological samples, thereby saving the step of comparing normal biological samples and effectively improving the efficiency of dose assessment. The first target gene involved in this embodiment includes: Net_sig,Ccnd1,Grn,Ifitm3,Phlda3,Ccng1,Sgta,Tcn2,Lrg1,Tmbim6,Xdh,Bax,Cd5l,Cd19,Glul,Spib,Swap70,Ccr7,Rhoc,Ly6d.

[0071] Step S2: Input the expression information of the first target gene into the preset dose assessment model so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene.

[0072] The construction of the dose assessment model includes:

[0073] Obtain several sets of training samples of subjects exposed to known doses of radiation and training samples of subjects not exposed to radiation;

[0074] Transcriptome sequencing was performed on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples.

[0075] The transcriptome sequencing information of the irradiated training samples was compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes.

[0076] By screening for the correlation of features of differentially expressed genes, a second target gene is obtained;

[0077] Quantitative gene amplification fluorescence detection was performed on the second target gene to obtain its expression information;

[0078] Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model.

[0079] In step S2, the expression information of the first target gene obtained in step S1 is input into a pre-defined dose assessment model. The model can then determine the radiation dose of the biological sample to be assessed based on this information. The dose assessment model is trained using the actual radiation dose of the irradiated training sample and the corresponding expression information of the second target gene as input, and the predicted radiation dose of the training sample as output. Through training, the dose assessment model learns the relationship between gene expression information and radiation dose, thus accurately predicting the radiation dose of a given biological sample given its gene expression information. This model can play an important role in the field of radiation dose assessment, helping scientists and doctors to more quickly assess the radiation dose of biological samples for further analysis and processing.

[0080] The training process of the dose assessment model is explained in detail below:

[0081] like Figure 2 As shown, the training process of the dose assessment model includes the following steps:

[0082] Step S201: Obtain several sets of training samples of training subjects who have been exposed to known doses of radiation and training samples of subjects who have not been exposed to radiation;

[0083] For step S201, multiple groups of irradiated training samples with known radiation doses are obtained, along with normal training samples from the same training subject that have not been irradiated. It is important to note that the normal training samples used as controls and the training samples with known radiation doses must come from the same training subject. This reduces the impact of individual genetic differences on the results, thus providing a better understanding of the relationship between gene expression information and radiation dose.

[0084] Step S202: Perform transcriptome sequencing on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples.

[0085] In step S202, transcriptome sequencing is performed on both the irradiated and unirradiated training samples in each training sample group. Transcriptome sequencing first involves extracting RNA from the samples and converting it into cDNA. Then, the cDNA is sequenced using a sequencer to obtain a large number of sequencing reads. Next, by comparing these reads with normal biological samples, the expression level of each gene can be determined. The expression levels of all genes in each training sample are measured to obtain the corresponding transcriptome sequencing information. This information includes numerical data on gene expression levels, which can be used to compare the differences in gene expression between irradiated and normal training samples. By analyzing this transcriptome sequencing data, we can obtain detailed information about the effects of radiation on gene expression.

[0086] Step S203: Compare and analyze the transcriptome sequencing information of the irradiated training samples with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes;

[0087] For step S203, the transcriptome sequencing information of the irradiated training samples obtained in step S202 is compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples. By comparing the gene expression levels of the two groups of samples, genes with significantly altered expression in the irradiated group compared with the unirradiated group can be identified, thus obtaining differentially expressed genes.

[0088] Step S204: Perform correlation screening on the differentially expressed genes to obtain the second target gene;

[0089] In a preferred embodiment, the differentially expressed genes are subjected to correlation screening of features to obtain a second target gene, including:

[0090] For each differentially expressed gene, calculate the logarithmic fold change, significance level, and correlation coefficient:

[0091] Differentially expressed genes that simultaneously meet the second, third, and fourth preset conditions are selected from the differentially expressed genes and used as the second target genes for preliminary screening. The second preset condition is: the absolute value of the logarithmic fold change is greater than the first preset threshold; the third preset condition is: the significance level is less than the second preset threshold; and the fourth preset condition is: the absolute value of the correlation coefficient is greater than or equal to the third preset threshold.

[0092] The Boruta algorithm was used to perform feature screening on the initially screened second target genes to obtain the final second target genes.

[0093] In step S204, by comparing the irradiated and unirradiated samples in step S203, genes whose transcriptome sequencing information differs due to radiation exposure are identified. However, to ensure the accuracy and reliability of model training, it is necessary to further screen these genes with different transcriptome sequencing information to identify differentially expressed genes that are strongly correlated with radiation dose.

[0094] First, statistical methods were used to initially screen for differentially expressed genes correlated with dosage. Specifically, the logarithmic fold change (log2 FC), significance level (p-value), and correlation coefficient were calculated for each gene among those with different transcriptome sequencing information. Next, differentially expressed genes that met the following preset thresholds were selected from these genes with different transcriptome sequencing information: |log2 FC| > 1, indicating at least a two-fold change in gene expression level; p-value < 0.05 to ensure significant differences in gene expression; and |Pearson correlation coefficient| ≤ 0.3, indicating a strong correlation between the differentially expressed gene and dosage.

[0095] After initial screening using statistical methods, the Boruta algorithm was further used for feature selection. The Boruta algorithm is a feature selection algorithm that comprehensively considers the correlation between genes and dosage and their importance in the random forest. By iteratively evaluating each differential gene, the importance index of the original gene is compared with the importance index of randomly generated features in the random forest model, and finally the second target gene with strong correlation to dosage is selected.

[0096] It should be noted that the first target gene used in actual radiation dose assessment is the same as the second target gene obtained through dose assessment model training; both represent genes that are strongly correlated with dose.

[0097] Step S205: Quantitative gene amplification fluorescence detection is performed on the second target gene to obtain the expression information of the second target gene;

[0098] In step S205, quantitative PCR (qPCR) is used to quantitatively measure the expression level of a specific second target gene. This is achieved by designing suitable primers and probes for the second target gene to amplify and detect it. During amplification, the primers or probes are labeled to emit a fluorescent signal. The expression information of the second target gene is then obtained by measuring the intensity of the fluorescence signal. Quantitative PCR of the second target gene provides quantitative data on its expression level.

[0099] Step S206: Using the actual radiation dose of the irradiated training sample and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, train the dose assessment model to be trained to obtain the trained dose assessment model.

[0100] In a preferred embodiment, the dose assessment model to be trained is trained using the actual radiation dose of the irradiated training sample and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, to obtain the trained dose assessment model, including:

[0101] The irradiated training samples are divided into a first training set and a first test set according to a preset ratio;

[0102] The actual radiation dose of each training sample in the first training set and the expression information of the corresponding second target gene are input into the dose assessment model to be trained for training until the number of training times exceeds the preset number, and the trained dose assessment model is obtained.

[0103] For step S206, the irradiated training samples are first divided into two parts according to a preset ratio (7:3): a first training set N_train and a first test set N_test. The training set N_train is used to train the dose assessment model, while the test set N_test is used to validate and evaluate the trained model.

[0104] Then, the first training set N_train is input into the dose assessment model to be trained. The expression information of each second target gene in the first training set N_train and the corresponding actual radiation dose are used as inputs, and the predicted radiation dose is used as outputs for training until the number of training times reaches the preset number. In this embodiment, the preset number of training times is 50.

[0105] The following is a detailed explanation of each training process of the dose assessment model:

[0106] In a preferred embodiment, the dose assessment model includes: a basic sub-model and an optimized sub-model;

[0107] The actual radiation dose of each training sample in the first training set and the corresponding expression information of the second target gene are input into the dose assessment model to be trained for training until the first training iteration exceeds the first preset iteration, resulting in a trained dose assessment model, including:

[0108] Divide the first training set into several training subsets;

[0109] During each training session, one training subset is selected sequentially from all training subsets as the second test set, and the remaining training subsets are used as the second training set; the training subsets can be selected cyclically.

[0110] The basic sub-model is trained based on the data in the second training set until the second training number exceeds the second preset number, so as to obtain the trained basic sub-model and the first predicted radiation dose obtained by each training of the basic sub-model.

[0111] Based on the first predicted radiation doses obtained from each training of the basic sub-model and the corresponding data of each first predicted radiation dose in the second training set, the optimized sub-model is trained until the third training number exceeds the third preset number, and the trained optimized sub-model is obtained.

[0112] The second test set is input into the trained base sub-model and the trained optimized sub-model to obtain the predicted radiation dose for each data in the second test set.

[0113] Based on the predicted radiation dose of each data in the second test set and the actual radiation dose corresponding to each data in the second test set, the values ​​of the coefficient of determination and root mean square error are calculated.

[0114] If the number of training iterations is less than the first preset number of iterations, then continue with the next training iteration; otherwise, compare the values ​​of the coefficient of determination and root mean square error obtained from each training iteration, and take the dose assessment model with the largest coefficient of determination and the smallest root mean square error as the completed dose assessment model.

[0115] In a preferred embodiment, the values ​​of the coefficient of determination and the root mean square error are calculated based on the predicted radiation dose of each data point in the second test set and the actual radiation dose corresponding to each data point in the second test set, including:

[0116] Based on the predicted radiation dose and the corresponding actual radiation dose for each data point in the second test set, the coefficient of determination is calculated using the following formula:

[0117]

[0118] Based on the predicted radiation dose and the corresponding actual radiation dose of each data point in the second test set, the root mean square error is calculated using the following formula:

[0119]

[0120] Among them, R 2 Y represents the coefficient of determination, RMSE represents the root mean square error, and Y represents the root mean square error. i This represents the actual radiation dose of the i-th data point in the second test set. This represents the predicted radiation dose for the i-th data point in the second test set. denoted as the average actual radiation dose in the second test set, and n represents the total amount of data in the test set.

[0121] In one embodiment of the present invention, the dose assessment model includes two parts: a basic sub-model and an optimized sub-model. The basic sub-model is used to establish the initial dose assessment model, while the optimized sub-model further integrates and corrects the prediction results based on the prediction results of the basic sub-model.

[0122] During the training of the dose assessment model, the first training set is divided into five training subsets, i.e., five-fold cross-validation. During each training iteration, one training subset is sequentially selected as the second test set N_train_test, and the remaining training subsets are used as the second training set N_train_train. This method yields multiple training and test combinations, increasing the number of samples for training the model and making more effective use of available data. These five training subsets can be selected cyclically to ensure the model receives comprehensive training.

[0123] In each training iteration, the expression information of each second target gene in the second training set N_train_train is first input into the base sub-model for training until the training iterations of the base sub-model meet the preset number of iterations. The optimal base sub-model is selected based on the root mean square error obtained from each training iteration, thus obtaining the trained base sub-model and its first predicted radiation dose for the data in the second training set N_train_train. Next, based on the first predicted radiation dose and the corresponding data in the second training set, the optimization sub-model is trained until the training iterations of the optimization sub-model meet the preset number of iterations. The optimal optimization sub-model is selected based on the root mean square error obtained from each training iteration, thus obtaining the trained base sub-model and optimization sub-model. Then, by inputting the second test set into the trained base sub-model and optimization sub-model, the prediction results of the base sub-model and optimization sub-model on the second test set are calculated. Based on the comparison between the prediction results and the actual radiation dose, the determination coefficient R of the base sub-model and optimization sub-model is calculated using a formula. 2 The root mean square error (RMSE) is used to determine whether the required number of training iterations has been reached, based on preset requirements for the coefficient of determination, the RMSE threshold, and the number of iterations. If the requirements are met, the coefficient of determination R0 obtained from each training iteration is used to determine whether the required number of training iterations has been reached. 2 By comparing the root mean square error (RMSE) with the values, 50 sets of determination coefficients R are obtained in this embodiment. 2 The root mean square error (RMSE) is used to select the coefficient of determination R. 2 The dose assessment model corresponding to the maximum and minimum root mean square error (RMSE) is used as the completed dose assessment model. If the number of training iterations does not meet the requirements, the next training iteration is performed, and the next training subset is selected as the second test set for the next training iteration, and training continues.

[0124] It should be noted that if there is no coefficient of determination R... 2 If the maximum simultaneous root mean square error (RMSE) is minimized, then the coefficient of determination R is selected according to a certain weight. 2 The dose assessment model corresponding to a larger value and a smaller RMSE.

[0125] In a preferred embodiment, the base sub-model includes a combination of the following regression sub-models: random forest regression sub-model, linear regression sub-model, XGBoost regression sub-model, CatBoost regression sub-model, elastic network regression sub-model, and support vector regression sub-model.

[0126] In a preferred embodiment, the base sub-model is trained based on data from the second training set until the second training iterations exceed a second preset number of iterations, resulting in a trained base sub-model and each first predicted radiation dose obtained from each training iteration of the base sub-model, including:

[0127] The basic sub-model is trained based on the data in the second training set; wherein, during each training, for each regression sub-model, the expression information of the second target gene of each data in the second training set is used as input and the first predicted radiation dose of each data in the second training set is used as output for training.

[0128] For each regression sub-model, the root mean square error of each regression sub-model is calculated based on the first predicted radiation dose and the corresponding actual radiation dose for each data in each second training set. The root mean square errors of each regression sub-model are then weighted to obtain the first loss function of the base sub-model.

[0129] Determine if the second number of training iterations is less than the second preset number of iterations. If so, repeat the next training iteration. If not, compare the first loss function values ​​obtained from each training iteration and use the base sub-model with the smallest first loss function value as the completed base sub-model.

[0130] In one embodiment of the present invention, the base sub-model is composed of a stacked combination of a random forest regression sub-model, a linear regression sub-model, and an XGBoost regression sub-model. Therefore, during the training process of the base sub-model, the expression information of each second target gene in the second training set N_train_train is input into the base sub-model. That is, the expression information of each second target gene in the second training set is input into the three regression sub-models respectively. For the expression information of each second target gene, it is used as input, and the random forest regression sub-model, the linear regression sub-model, and the XGBoost regression sub-model are used for training respectively, and the predicted radiation dose results of the three regression sub-models are obtained. n represents the expression information of the nth second target gene in the second training set, and i represents the i-th training of the basic sub-model.

[0131] Each regression sub-model has its own learning algorithm and feature representation capabilities. By inputting the expression information of the second target gene into these models, their diversity can be utilized to explore different patterns and associations in the data, yielding different results. The Random Forest regression sub-model has good robustness and generalization ability, and can effectively handle nonlinear relationships between multiple input features; the Linear Regression sub-model is suitable for handling data with strong linear correlations and can provide good interpretability; while the XGBoost regression sub-model is a powerful gradient boosting-based model with high predictive performance when dealing with complex data and nonlinear relationships.

[0132] For each regression sub-model, the calculated predicted radiation dose is compared with the corresponding actual radiation dose, and the root mean square error (RMSE) of each regression sub-model is calculated. The weighted value of the RMSE of each regression sub-model is used as the loss function. When the training times of the base sub-model meet the preset training times, the first loss function obtained from each training session is compared, and the base sub-model corresponding to the minimum value of the first loss function is taken as the trained base sub-model. Each first predicted radiation dose obtained by this base sub-model in each training session is output. If the preset number of training iterations is not reached, the training will be repeated on the second training set.

[0133] In a preferred embodiment, the optimized sub-model is trained based on the first predicted radiation doses obtained from each training iteration of the base sub-model and the corresponding data for each first predicted radiation dose in the second training set, until the third training iteration exceeds a third preset iteration, resulting in a trained optimized sub-model, including:

[0134] Each first predicted radiation dose obtained from each training of the basic sub-model is divided into several groups of predicted data; among them, the predicted data includes: each first predicted radiation dose obtained from the same training of the basic sub-model.

[0135] The optimized sub-model is trained based on the first predicted radiation doses obtained from each training of the base sub-model and the corresponding data of each first predicted radiation dose in the second training set. In each training session, a set of predicted data is selected sequentially from several sets of predicted data. For each first predicted radiation dose in the set of predicted data and the corresponding data of the first predicted radiation dose in the second training set, the training is performed with the first predicted radiation dose and the corresponding data of the first predicted radiation dose in the second training set as inputs and the second predicted radiation dose of the second training set data as outputs.

[0136] The second loss function is calculated based on the second predicted radiation dose and the corresponding actual radiation dose for each second training set data.

[0137] Determine if the third training iteration is less than the third preset iteration. If so, continue with the next training iteration. If not, compare the second loss function values ​​obtained from each training iteration and select the optimized sub-model with the smallest second loss function value as the completed optimized sub-model.

[0138] In one embodiment of the present invention, the result obtained by the basic sub-model is the predicted radiation dose result based on three independent regression sub-models. However, we wish to further optimize these results and integrate them to obtain a more accurate prediction result. To this end, each first predicted radiation dose output by the basic sub-model is combined. The first predicted radiation dose is used as input, along with the expression information of the second target gene corresponding to the second training set. Each time j takes the values ​​1, 2, 3, ..., i, the second predicted radiation dose from the second training set data is... Training is performed for the output; the second predicted radiation dose is obtained from this training. By comparing the actual radiation dose, the root mean square error of this training can be calculated and used as the second loss function to determine whether the training times of the optimized sub-model have reached the preset number. If the optimized sub-model has reached the preset number of training times, the second loss function of each training session is obtained. The second loss function values ​​obtained from each training session are compared, and the optimized sub-model corresponding to the smallest second loss function value is selected as the optimized sub-model that has been trained. The smallest second loss function value means that the optimized sub-model has reached a good training state.

[0139] It should be noted that the optimization sub-model aims to integrate the advantages of different models and combine the results of each regression sub-model obtained from the basic sub-model to obtain a final prediction result.

[0140] Finally, by comparing the determination coefficient and root mean square error values ​​obtained from each training iteration, the optimal dose assessment model was selected as the trained dose assessment model. The trained dose assessment model was then used to assess the radiation dose on the first test set N_test, and the results are as follows: Figure 3 As shown, the coefficient of determination R 2 The coefficient of determination (R²) is 0.852, which is greater than 85%. The R² is an indicator that assesses how well a model fits the data. Its value ranges from 0 to 1; the closer it is to 1, the stronger the model's explanatory power. In this case, the R² is... 2The mean square error (RMSE) was 0.852, exceeding the 85% threshold, indicating that the model can well explain the variance of the test set data and has high predictive ability. The root mean square error (RMSE) was 1.292 Gy, less than 1.5 Gy. RMSE is an indicator that measures the difference between predicted and true values; the smaller the value, the higher the prediction accuracy. In this case, the RMSE of 1.292 Gy, less than the 1.5 Gy threshold, indicates that the model has high prediction accuracy and the difference between the predicted result and the true value is relatively small. In summary, the trained dose assessment model performed well, with high predictive ability and accuracy.

[0141] like Figure 4 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided;

[0142] One embodiment of the present invention provides a biological radiation dose assessment device, comprising: a first target gene expression information extraction module and a dose assessment module;

[0143] The first target gene expression information extraction module is used to extract the expression information of the first target gene in the biological sample to be evaluated;

[0144] The dose assessment module is used to input the expression information of the first target gene into a preset dose assessment model so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene.

[0145] The dose assessment module includes: a training sample acquisition submodule, a transcriptome sequencing submodule, a differential gene analysis submodule, a differential gene screening submodule, a gene expression information detection submodule, and a dose assessment model training submodule.

[0146] The training sample acquisition submodule is used to acquire several sets of training samples of training objects subjected to known radiation doses and training samples of those not subjected to radiation.

[0147] The transcriptome sequencing submodule is used to perform transcriptome sequencing on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples.

[0148] The differential gene analysis submodule is used to compare and analyze the transcriptome sequencing information of irradiated training samples with the transcriptome sequencing information of corresponding unirradiated training samples to obtain differential genes.

[0149] The differential gene screening submodule is used to screen differential genes for relevance of features to obtain the second target gene;

[0150] The gene expression information detection submodule is used to perform quantitative gene amplification fluorescence detection on the second target gene to obtain the expression information of the second target gene;

[0151] The dose assessment model training submodule is used to train the dose assessment model to be trained by taking the actual radiation dose of the irradiated training sample and the expression information of the corresponding second target gene as input and the predicted radiation dose as output, so as to obtain the trained dose assessment model.

[0152] In a preferred embodiment, the differential gene screening submodule includes: a differential expression calculation unit, a first screening unit, and a second screening unit;

[0153] The differential expression calculation unit is used to calculate the logarithmic fold change, significance level, and correlation coefficient for each differentially expressed gene.

[0154] The first screening unit is used to screen differentially expressed genes that simultaneously meet the second, third, and fourth preset conditions, and use them as the second target genes for preliminary screening. The second preset condition is: the absolute value of the logarithmic fold change is greater than the first preset threshold; the third preset condition is: the significance level is less than the second preset threshold; and the fourth preset condition is: the absolute value of the correlation coefficient is greater than or equal to the third preset threshold.

[0155] The second screening unit is used to perform feature screening on the second target gene in the initial screening using the Boruta algorithm to obtain the second target gene.

[0156] In a preferred embodiment, the dose assessment model training submodule includes: a first data partitioning unit and a training unit;

[0157] The first data partitioning unit is used to divide the irradiated training samples into a first training set and a first test set according to a preset ratio;

[0158] The training unit is used to input the actual radiation dose of each training sample in the first training set and the expression information of the corresponding second target gene into the dose assessment model to be trained for training until the number of training times exceeds the preset number, and the trained dose assessment model is obtained.

[0159] In a preferred embodiment, the training unit includes: a second data partitioning subunit, a second training set selection subunit, a basic sub-model training subunit, an optimized sub-model training subunit, a second test set prediction subunit, a model performance calculation subunit, and a first training iteration discrimination subunit.

[0160] The second data partitioning subunit is used to divide the first training set into several training subsets;

[0161] The second training set selection subunit is used to select one training subset from all training subsets in sequence as the second test set during each training session, and the remaining training subsets as the second training set; the training subsets can be selected cyclically.

[0162] The basic sub-model training sub-unit is used to train the basic sub-model based on the data in the second training set until the second training number exceeds the second preset number, so as to obtain the trained basic sub-model and the first predicted radiation dose obtained by each training of the basic sub-model.

[0163] The optimized sub-model training sub-unit is used to train the optimized sub-model based on the first predicted radiation doses obtained from each training of the basic sub-model and the data corresponding to each first predicted radiation dose in the second training set, until the third training number exceeds the third preset number, and the trained optimized sub-model is obtained.

[0164] The second test set prediction sub-unit is used to input the second test set into the trained base sub-model and the trained optimization sub-model to obtain the predicted radiation dose of each data in the second test set.

[0165] The model performance calculation subunit is used to calculate the values ​​of the coefficient of determination and the root mean square error based on the predicted radiation dose of each data in the second test set and the actual radiation dose corresponding to each data in the second test set.

[0166] The first training iteration discrimination subunit is used to determine whether the first training iteration is less than the first preset iteration. If so, the next training iteration continues; otherwise, the values ​​of the determination coefficient and root mean square error obtained from each training iteration are compared, and the dose assessment model with the largest determination coefficient and the smallest root mean square error is taken as the completed dose assessment model.

[0167] In a preferred embodiment, the basic sub-model training sub-unit includes: a basic sub-model iterative training component, a first loss function calculation component, and a second training iteration discrimination component;

[0168] The basic sub-model iterative training component is used to train the basic sub-model based on the data in the second training set. In each training session, for each regression sub-model, the expression information of the second target gene of each data in the second training set is used as input, and the first predicted radiation dose of each data in the second training set is used as output for training.

[0169] The first loss function calculation component is used to calculate the root mean square error of each regression sub-model based on the first predicted radiation dose and the corresponding actual radiation dose of each data in each second training set for each regression sub-model, and to calculate the first loss function of the basic sub-model by weighting the root mean square errors of each regression sub-model.

[0170] The second training iteration discrimination component is used to determine whether the second training iteration is less than the second preset iteration. If it is, the next training iteration is repeated; if not, the first loss function value obtained from each training iteration is compared, and the base sub-model with the smallest first loss function value is taken as the base sub-model that has been trained.

[0171] In a preferred embodiment, the optimized sub-model training sub-unit includes: a prediction data grouping component, an optimized sub-model iterative training component, a second loss function calculation component, and a third training iteration discrimination component;

[0172] The prediction data grouping component is used to divide each first predicted radiation dose obtained from each training of the base sub-model into several groups of prediction data; wherein, the prediction data includes: each first predicted radiation dose obtained from the same training of the base sub-model.

[0173] An optimized sub-model iterative training component is used to train the optimized sub-model based on each first predicted radiation dose obtained from each training of the base sub-model and the corresponding data of each first predicted radiation dose in the second training set. Specifically, during each training, a set of predicted data is selected sequentially from several sets of predicted data. For each first predicted radiation dose in the set of predicted data and the corresponding data of the first predicted radiation dose in the second training set, training is performed with the first predicted radiation dose and the corresponding data of the first predicted radiation dose in the second training set as input and the second predicted radiation dose of the second training set data as output.

[0174] The second loss function calculation component is used to calculate the second loss function based on the second predicted radiation dose and the corresponding actual radiation dose for each second training set data.

[0175] The third training iteration discrimination component is used to determine whether the third training iteration is less than the third preset iteration. If it is, the next training iteration continues; if not, the second loss function value obtained from each training iteration is compared, and the optimized sub-model with the smallest second loss function value is taken as the optimized sub-model that has been trained.

[0176] In a preferred embodiment, the model performance calculation subunit includes: a coefficient of determination calculation component and a root mean square error calculation component;

[0177] The coefficient of determination calculation component is used to calculate the coefficient of determination based on the predicted radiation dose and the corresponding actual radiation dose of each data point in the second test set, using the following formula:

[0178]

[0179] The root mean square error (RMSE) calculation component is used to calculate the RMSE based on the predicted radiation dose and the corresponding actual radiation dose of each data point in the second test set, using the following formula:

[0180]

[0181] Among them, R 2 Y represents the coefficient of determination, RMSE represents the root mean square error, and Y represents the root mean square error. i This represents the actual radiation dose of the i-th data point in the second test set. This represents the predicted radiation dose for the i-th data point in the second test set. denoted as the average actual radiation dose in the second test set, and n represents the total amount of data in the test set.

[0182] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the biological radiation dose assessment method provided by any of the above-described method embodiments of the present invention.

[0183] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0184] Based on the above embodiments of the biological radiation dose assessment method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the biological radiation dose assessment method of any embodiment of the present invention.

[0185] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0186] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0187] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0188] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for assessing biological radiation dose, characterized in that, include: Extract the expression information of the first target gene from the biological sample to be evaluated; The expression information of the first target gene is input into a preset dose assessment model so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene. The construction of the dose assessment model includes: Obtain several sets of training samples of subjects subjected to known radiation doses and training samples of subjects not subjected to radiation. Transcriptome sequencing was performed on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples. The transcriptome sequencing information of the irradiated training samples was compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes. The process involves: screening differentially expressed genes for relevance features to obtain second target genes; this screening includes: for each differentially expressed gene, calculating its logarithmic fold change, significance level, and correlation coefficient; selecting differentially expressed genes that simultaneously meet a second, third, and fourth preset condition as preliminary second target genes; the second preset condition being: the absolute value of the logarithmic fold change is greater than a first preset threshold; the third preset condition being: the significance level is less than the second preset threshold; and the fourth preset condition being: the absolute value of the correlation coefficient is greater than or equal to the third preset threshold; and using the Boruta algorithm to perform feature screening on the preliminary second target genes to obtain the final second target genes. The expression information of the second target gene was obtained by quantitative gene amplification fluorescence detection. Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model.

2. The biological radiation dose assessment method according to claim 1, characterized in that, Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model, including: The irradiated training samples are divided into a first training set and a first test set according to a preset ratio; The actual radiation dose of each training sample in the first training set and the expression information of the corresponding second target gene are input into the dose assessment model to be trained for training until the number of training times exceeds the preset number, and the trained dose assessment model is obtained.

3. The biological radiation dose assessment method according to claim 2, characterized in that, The dose assessment model includes: a basic sub-model and an optimized sub-model; The step of inputting the actual radiation dose of each training sample in the first training set and the corresponding expression information of the second target gene into the dose assessment model to be trained for training, until the first training iteration exceeds the first preset iteration, to obtain the trained dose assessment model, includes: Divide the first training set into several training subsets; During each training session, one training subset is selected sequentially from all training subsets as the second test set, and the remaining training subsets are used as the second training set; the training subsets can be selected cyclically. The basic sub-model is trained based on the data in the second training set until the second training number exceeds the second preset number, so as to obtain the trained basic sub-model and the first predicted radiation dose obtained by each training of the basic sub-model. Based on the first predicted radiation doses obtained from each training of the basic sub-model and the corresponding data of each first predicted radiation dose in the second training set, the optimized sub-model is trained until the third training number exceeds the third preset number, and the trained optimized sub-model is obtained. The second test set is input into the trained base sub-model and the trained optimized sub-model to obtain the predicted radiation dose for each data in the second test set. Based on the predicted radiation dose of each data in the second test set and the actual radiation dose corresponding to each data in the second test set, the values ​​of the coefficient of determination and root mean square error are calculated. If the number of training iterations is less than the first preset number of iterations, then continue with the next training iteration; otherwise, compare the values ​​of the coefficient of determination and root mean square error obtained from each training iteration, and take the dose assessment model with the largest coefficient of determination and the smallest root mean square error as the completed dose assessment model.

4. The biological radiation dose assessment method according to claim 3, characterized in that, The basic sub-model includes a combination of the following regression sub-models: random forest regression sub-model, linear regression sub-model, XGBoost regression sub-model, CatBoost regression sub-model, elastic network regression sub-model, and support vector regression sub-model.

5. The biological radiation dose assessment method according to claim 4, characterized in that, The base sub-model is trained based on the data in the second training set until the second training iterations exceed the second preset number of iterations, resulting in the trained base sub-model and the first predicted radiation doses obtained from each training iteration of the base sub-model, including: The basic sub-model is trained based on the data in the second training set; wherein, during each training, for each regression sub-model, the expression information of the second target gene of each data in the second training set is used as input and the first predicted radiation dose of each data in the second training set is used as output for training. For each regression sub-model, the root mean square error of each regression sub-model is calculated based on the first predicted radiation dose and the corresponding actual radiation dose for each data in each second training set. The root mean square errors of each regression sub-model are then weighted to obtain the first loss function of the base sub-model. Determine if the second number of training iterations is less than the second preset number of iterations. If so, repeat the next training iteration. If not, compare the first loss function values ​​obtained from each training iteration and use the base sub-model with the smallest first loss function value as the completed base sub-model.

6. The biological radiation dose assessment method according to claim 5, characterized in that, Based on the first predicted radiation doses obtained from each training iteration of the base sub-model and the corresponding data in the second training set, the optimized sub-model is trained until the third training iteration exceeds the third preset number of iterations, resulting in a fully trained optimized sub-model, including: Each first predicted radiation dose obtained from each training of the basic sub-model is divided into several groups of predicted data; wherein, the predicted data includes: each first predicted radiation dose obtained from the same training of the basic sub-model. The optimized sub-model is trained based on the first predicted radiation doses obtained from each training of the base sub-model and the corresponding data of each first predicted radiation dose in the second training set. In each training session, a set of predicted data is selected sequentially from several sets of predicted data. For each first predicted radiation dose in the set of predicted data and the corresponding data of the first predicted radiation dose in the second training set, the training is performed with the first predicted radiation dose and the corresponding data of the first predicted radiation dose in the second training set as inputs and the second predicted radiation dose of the second training set data as outputs. The second loss function is calculated based on the second predicted radiation dose and the corresponding actual radiation dose for each second training set data. Determine if the third training iteration is less than the third preset iteration. If so, continue with the next training iteration. If not, compare the second loss function values ​​obtained from each training iteration and select the optimized sub-model with the smallest second loss function value as the completed optimized sub-model.

7. The biological radiation dose assessment method according to claim 6, characterized in that, Based on the predicted radiation dose of each data point in the second test set and the corresponding actual radiation dose of each data point in the second test set, the values ​​of the coefficient of determination and root mean square error are calculated, including: Based on the predicted radiation dose and the corresponding actual radiation dose of each data point in the second test set, the coefficient of determination is calculated using the following formula: ； Based on the predicted radiation dose and the actual radiation dose corresponding to each data point in the second test set, the root mean square error is calculated using the following formula: ； in, The coefficient of determination is represented by the coefficient of determination. This represents the root mean square error. Indicates the second test set The actual radiation dose of each data point. Indicates the second test set Predicted radiation dose from data points This represents the average actual radiation dose in the second test set. This indicates the total amount of data in the test set.

8. A biological radiation dose assessment device, characterized in that, include: The first target gene expression information extraction module and the dose assessment module; The first target gene expression information extraction module is used to extract the expression information of the first target gene in the biological sample to be evaluated; The dose assessment module is used to input the expression information of the first target gene into a preset dose assessment model, so that the dose assessment model can determine the radiation dose of the biological sample to be assessed based on the expression information of the first target gene. The construction of the dose assessment model includes: Obtain several sets of training samples of subjects subjected to known radiation doses and training samples of subjects not subjected to radiation. Transcriptome sequencing was performed on each group of training samples to obtain the transcriptome sequencing information of the irradiated training samples and the corresponding unirradiated training samples. The transcriptome sequencing information of the irradiated training samples was compared and analyzed with the transcriptome sequencing information of the corresponding unirradiated training samples to obtain differentially expressed genes. The process involves: screening differentially expressed genes for relevance features to obtain second target genes; this screening includes: for each differentially expressed gene, calculating its logarithmic fold change, significance level, and correlation coefficient; selecting differentially expressed genes that simultaneously meet a second, third, and fourth preset condition as preliminary second target genes; the second preset condition being: the absolute value of the logarithmic fold change is greater than a first preset threshold; the third preset condition being: the significance level is less than the second preset threshold; and the fourth preset condition being: the absolute value of the correlation coefficient is greater than or equal to the third preset threshold; and using the Boruta algorithm to perform feature screening on the preliminary second target genes to obtain the final second target genes. The expression information of the second target gene was obtained by quantitative gene amplification fluorescence detection. Using the actual radiation dose of the irradiated training samples and the expression information of the corresponding second target gene as input, and the predicted radiation dose as output, the dose assessment model to be trained is trained to obtain the trained dose assessment model.

9. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the biological radiation dose assessment method as described in any one of claims 1-7.

Images