A circRNA marker combination for radiation dose classification and application thereof

Abstract

The application discloses a circRNA marker combination for radiation dose classification and application thereof, and belongs to the technical field of biotechnology and nuclear radiation protection technology. In view of the technical problem that existing radiation markers have a short half-life and are not suitable for dose classification of samples collected at a late time point after a radiation accident, the application discloses a circRNA marker combination for radiation dose classification, which is composed of Hsa_circZDHHC21_004, Hsa_circATP5C1_006 and Hsa_circMPP6_020. The three kinds of circRNA are significantly down-regulated in a dose-dependent manner in peripheral blood 48-72 hours after irradiation, and the expression is not affected by gender and age. Based on the expression level, a model established by using a support vector machine can effectively distinguish whether irradiation is performed and 2 Gy or not, has a high accuracy, and is suitable for rapid dose classification of samples collected at a late time point of a radiation accident.

Description

Technical Field

[0001] This invention belongs to the fields of biotechnology and nuclear radiation protection technology, specifically relating to a combination of circRNA biomarkers for radiation dose classification and their applications. Background Technology

[0002] The severity of damage from acute whole-body irradiation can be categorized into mild (<2 Gy), moderate (2-4 Gy), severe (4-6 Gy), very severe (6-8 Gy), and lethal (>8 Gy) depending on the radiation dose received. In accidents involving large-scale exposure to ionizing radiation, such as the Chernobyl and Fukushima nuclear accidents, quickly identifying exposed individuals and their dose ranges allows for the prediction of individual damage and the implementation of effective medical treatment. This is crucial for the rational allocation of medical resources and for saving the lives of those exposed in such accidents. Therefore, it is necessary to develop rapid, high-throughput radiation biomarkers for the rapid classification of radiation doses received by irradiated individuals.

[0003] Traditional genetic radiation biodosimeters, such as those using "dicentric chromosomes with loops," are not suitable for rapid dose classification in large populations due to the time-consuming sample preparation and result analysis. In recent years, some molecular biological markers, such as mRNA, miRNA, and proteins, have been found to exhibit dose-dependent changes in expression levels after ionizing radiation induction. However, these molecular biological markers typically have short half-lives and are relatively unstable in body fluids. If body fluid samples are collected from victims several days after a radiation accident, using these short-lived markers may not accurately determine the radiation dose range. Therefore, it is necessary to continue searching for suitable radiation biomarkers suitable for rapid dose classification in large populations.

[0004] Circular RNA (circRNA) is a novel class of non-coding RNA formed by back-splicing of precursor mRNA of coding genes. Its molecular structure is a covalently closed circular shape without free ends, making it more resistant to digestion by the nuclease RNase R compared to linear RNA. Studies have confirmed that the half-life of circRNA is significantly longer than that of the linear transcripts of its source genes. Enuka et al. calculated the half-lives of 60 circRNAs and their corresponding linear transcripts and found that the median half-life of circRNAs in mammalian cells (18.8–23.7 h) was at least 2.5 times that of their corresponding linear transcripts (4.0–7.4 h) (Enuka Y, Lauriola M, Feldman ME, Sas-Chen A, Ulitsky I, Yarden Y. 2015. Circular RNAs are long-lived and display only minimal early alterations in response to a growth factor. Nucleic Acids Res. 44(3): 1370–1383. doi:10.1093 / nar / gkv1367). The half-lives of CircHIPK3, circKIAA0182, circASXL1, and circLPAR1 can exceed 48 h, while the half-lives of their corresponding linear transcripts are less than 20 h (Jeck WR, Sorrentino JA, Wang K, Slevin MK, Burd CE, Liu J, Marzluff WF, Sharpless NE. 2013. Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA.19(2):141–157. doi:10.1261 / rna.035667.112). Therefore, circRNAs are a class of molecules that can exist relatively stably, giving them an advantage as radiation biomarkers.

[0005] circRNAs have been extensively studied as diagnostic biomarkers for diseases. Most studies have reported that circRNAs can serve as diagnostic and prognostic biomarkers for cancers such as gastric, esophageal, and liver cancer. They can also be used as diagnostic biomarkers for non-cancerous diseases such as cardiovascular, neurological, nephrological, and autoimmune diseases. It is known that ionizing radiation can alter the expression profile of circRNAs in cells and tissues, and that circRNA expression levels can be rapidly quantified; however, whether circRNAs can be used as radiation biomarkers for dose classification remains unclear. Summary of the Invention

[0006] To address the technical problems of existing radiation biomarkers having short half-lives and instability in body fluids, making them unsuitable for rapid dose classification of samples collected late after a radiation accident, this invention discloses a circRNA biomarker combination for radiation dose classification, consisting of Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020. These three circRNAs exhibit significant dose-dependent downregulation in peripheral blood 48-72 hours after exposure, and their expression is unaffected by gender or age. Based on their expression levels, a support vector machine model can effectively distinguish between exposure and doses around 2 Gy, demonstrating high accuracy at both 48 and 72 hours. This biomarker combination has a long half-life and good stability, making it suitable for rapid dose classification of samples collected late after a radiation accident, providing a powerful tool for large-scale population radiation damage assessment and medical decision-making.

[0007] To solve the above-mentioned technical problems and achieve the corresponding technical effects, the present invention provides the following technical solution: The first objective of this invention is to provide a circRNA biomarker set for radiation dose classification, the circRNA biomarker set consisting of Hsa_circZDHHC21_004, Hsa_circATP5C1_006 and Hsa_circMPP6_020, the nucleotide sequences of Hsa_circZDHHC21_004, Hsa_circATP5C1_006 and Hsa_circMPP6_020 being shown in SEQ ID NO.1, SEQ ID NO.2 and SEQ ID NO.3, respectively.

[0008] A second objective of this invention is to provide the application of the above-described combination of circRNA biomarkers in the preparation of products for classifying radiation doses of irradiated individuals.

[0009] In one embodiment of the present invention, the radiation dose classification also requires the collection of peripheral blood samples from the irradiated person, which are collected 48-72 hours after irradiation.

[0010] A third objective of this invention is to provide a reagent for detecting the expression levels of the above-mentioned circRNA biomarker combinations.

[0011] In one embodiment of the present invention, the reagent is used to detect the expression level of each circRNA in the circRNA biomarker combination.

[0012] In one embodiment of the present invention, the reagent comprises primer pairs specifically for detecting Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020.

[0013] A fourth object of the present invention is to provide the application of the above-described reagent in the preparation of a kit for classifying radiation doses in irradiated persons.

[0014] In one embodiment of the present invention, the radiation dose classification also requires the collection of peripheral blood samples from the irradiated person, which are collected 48-72 hours after irradiation.

[0015] A fifth object of the present invention is to provide a kit containing the above-described reagents for classifying radiation doses in irradiated individuals.

[0016] In one embodiment of the present invention, the radiation dose classification also requires the collection of peripheral blood samples from the irradiated person, which are collected 48-72 hours after irradiation.

[0017] The beneficial effects of this invention are: (1) Breaking through the bottleneck of dose assessment in late-stage sampling of radiation accidents The circRNA biomarker combination (Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020) provided by this invention still showed significant and stable dose-dependent downregulation of expression 48-72 h after irradiation. Its long half-life and high stability overcome the technical defects of traditional mRNA, miRNA and other molecular biomarkers, which cannot accurately classify dose at a later time point (such as 48-72 h) after radiation accident due to their short half-life and easy degradation in body fluids. It fills the technical gap in rapid dose assessment in the late stage of radiation accident.

[0018] (2) The model has the ability to distinguish dose thresholds, which accurately meets the needs of clinical treatment. The classification model built on the support vector machine algorithm in this invention can not only effectively distinguish whether someone has been exposed to radiation, but also has the ability to identify the key clinical threshold between dose levels above and below 2 Gy. In medical emergencies of nuclear and radiation accidents, 2 Gy is an important dividing line for deciding whether to initiate professional treatment: individuals exposed to doses below 2 Gy usually exhibit mild, reversible damage to the hematopoietic system, which can recover spontaneously under close observation, and are generally not considered priority patients; while when the exposure dose exceeds 2 Gy, patients may develop acute radiation sickness (such as myelopathic acute radiation sickness), requiring timely systemic medical intervention.

[0019] The high accuracy of this invention's model at the 2 Gy threshold enables emergency responders to quickly and accurately classify irradiated individuals into "priority treatment group" (>2 Gy) and "observable group" (≤2 Gy) within 48-72 hours after an accident (a time window when traditional biodosimeters are no longer effective). This allows for precise allocation of limited medical resources, preventing resource dilution and delays in the treatment of critically injured patients. Furthermore, accurately identifying individuals irradiated at 2 Gy and intervening early is of significant clinical importance in mitigating irreversible damage to the hematopoietic system, reducing the incidence of multiple organ failure, and improving the long-term quality of life for injured individuals.

[0020] (3) The biomarker expression is universal and applicable to large-scale population screening. The expression levels of the circRNA biomarker combination provided by this invention are not affected by gender or age, and are applicable to radiation-exposed populations of different genders and ages. It does not require complex corrections for individual characteristics and has good population universality and clinical operability.

[0021] (4) The model has high prediction accuracy and good time stability. The dose classification model established by the support vector machine algorithm in this invention shows excellent prediction accuracy at 48 h and 72 h after exposure. Furthermore, the model's recognition capability is further improved as the exposure time increases (72 h is better than 48 h), which can meet the emergency needs of rapid and high-throughput dose classification in large-scale radiation events such as nuclear accidents.

[0022] (5) Samples are easy to obtain and are suitable for large-scale emergency applications. This invention requires only peripheral blood samples, making sample acquisition convenient and minimally invasive. The testing process is fast and efficient, and it is highly compatible with existing clinical testing systems. It provides a reliable technical means for early injury prediction, triage and classification treatment, and rational allocation of medical resources for irradiated personnel in nuclear accidents and other ionizing radiation events. Attached Figure Description

[0023] Figure 1 Radiation dose-response and time-response characteristics of the expression levels of three circRNAs in human peripheral blood; among them... Figure 1 Figure A shows the radiation dose-response and time-response characteristics of Hsa_circZDHHC21_004 expression level in human peripheral blood. Figure 1 Figure B shows the radiation dose-response and time-response characteristics of Hsa_circATP5C1_006 expression level in human peripheral blood. Figure 1 C in the figure represents the radiation dose-response and time-response characteristics of Hsa_circMPP6_020 expression levels in human peripheral blood; data are expressed as mean ± standard deviation and analyzed using one-way ANOVA (*). P <0.05,** P <0.01, *** P <0.001, n=30); Figure 2 Radiation dose-response and time-response characteristics of the expression levels of seven circRNAs in human peripheral blood; among them... Figure 2 Figure A shows the radiation dose-response and time-response characteristics of Hsa_circPCMTD1_004 expression level in human peripheral blood. Figure 2 Figure B shows the radiation dose-response and time-response characteristics of Hsa_circSPECC1_006 expression level in human peripheral blood. Figure 2 C in the figure represents the radiation dose response and time response characteristics of Hsa_circFBXW7_009 expression level in human peripheral blood. Figure 2 The image shows the radiation dose-response and time-response characteristics of Hsa_circBARD1_018 expression levels in human peripheral blood, represented by D. Figure 2 E in the figure represents the radiation dose-response and time-response characteristics of Hsa_circFAM13B_024 expression level in human peripheral blood. Figure 2 F in the figure represents the radiation dose response and time response characteristics of Hsa_circZFAND6_008 expression level in human peripheral blood. Figure 2 G in the figure represents the radiation dose-response and time-response characteristics of Hsa_circXPO1_021 expression levels in human peripheral blood; data are expressed as mean ± standard deviation and analyzed by one-way ANOVA (*). P <0.05,** P <0.01, *** P <0.001, n=30); Figure 3 The figure shows the effect of gender on the expression levels of Hsa_circZDHHC21_004, Hsa_circATP5C1_006 and Hsa_circMPP6_020 in human peripheral blood at different dose points and time points after irradiation. Figure 3Figure AC in the figure shows the effect of sex on the expression levels of three circRNAs at different dose points 24 h after irradiation. Figure 3 The figure in DF shows the effect of sex on the expression levels of three circRNAs at different dose points 48 h after irradiation. Figure 3 The GI in the figure represents the effect of sex on the expression levels of three circRNAs at different dose points 72 h after irradiation; data are expressed as mean ± standard deviation, and one-way ANOVA was used, n=30; Figure 4 The figure shows the effect of age on the expression levels of Hsa_circZDHHC21_004, Hsa_circATP5C1_006 and Hsa_circMPP6_020 in human peripheral blood at different dose points and time points after irradiation. Figure 4 Figure AC in the figure shows the effect of age on the expression levels of three circRNAs at different dose points 24 h after irradiation. Figure 4 The diagram in DF shows the effect of age on the expression levels of three circRNAs at different dose points 48 h after irradiation. Figure 4 The GI in the figure represents the effect of age on the expression levels of three circRNAs at different dose points 72 h after irradiation; data are expressed as mean ± standard deviation, and one-way ANOVA was used, n=30; Figure 5 ROC curves for four dose classification models using support vector machines were generated; among them, Figure 5 In the figure, A represents the ROC curve distinguishing between the unirradiated and irradiated models 72 hours after irradiation. Figure 5 B in the graph represents the ROC curves for distinguishing between the 0–2 Gy and >2 Gy models 72 h after irradiation. Figure 5 In the figure, C represents the ROC curve distinguishing between the unirradiated and irradiated models 48 hours after irradiation. Figure 5 D in the figure represents the ROC curves of the 0~2 Gy and >2 Gy models 48 h after irradiation. Figure 6 ROC curves for four dose classification models of random forest are shown; among them, Figure 6 In the figure, A represents the ROC curve distinguishing between the unirradiated and irradiated models 72 hours after irradiation. Figure 6 B in the graph represents the ROC curves for distinguishing between the 0–2 Gy and >2 Gy models 72 h after irradiation. Figure 6 In the figure, C represents the ROC curve distinguishing between the unirradiated and irradiated models 48 hours after irradiation. Figure 6 D in the figure represents the ROC curves for distinguishing between the 0~2 Gy and >2 Gy models 48 h after irradiation. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings. It should be noted that the embodiments mentioned below are only for explaining the invention and are not intended to limit the scope of the invention. The embodiments mentioned below are only some embodiments of the invention, not all embodiments. Those skilled in the art can refer to the content of this document and appropriately improve the process parameters to achieve the objectives of the invention. It should be particularly noted that all similar substitutions and modifications are obvious to those skilled in the art and are considered to be included in the invention. The methods and applications of this invention have been described through preferred embodiments, and those skilled in the art can obviously modify or appropriately change and combine the methods and applications described herein without departing from the content and scope of this invention to realize and apply the technology of this invention. In the art, embodiments obtained by other those skilled in the art without creative effort are all protected by this invention.

[0025] Unless otherwise specified, the experimental methods used in the following examples are conventional methods, and the materials, reagents and instruments used are conventional materials, reagents and instruments in the art, which can be obtained by those skilled in the art through commercial channels.

[0026] The nucleotide sequences of all circRNAs involved in this invention are as follows: the nucleotide sequence of Hsa_circZDHHC21_004 is shown in SEQ ID NO.1, the nucleotide sequence of Hsa_circATP5C1_006 is shown in SEQ ID NO.2, the nucleotide sequence of Hsa_circMPP6_020 is shown in SEQ ID NO.3, the nucleotide sequence of Hsa_circPCMTD1_004 is shown in SEQ ID NO.4, the nucleotide sequence of Hsa_circSPECC1_006 is shown in SEQ ID NO.5, the nucleotide sequence of Hsa_circFBXW7_009 is shown in SEQ ID NO.6, the nucleotide sequence of Hsa_circBARD1_018 is shown in SEQ ID NO.7, the nucleotide sequence of Hsa_circFAM13B_024 is shown in SEQ ID NO.8, the nucleotide sequence of Hsa_circZFAND6_008 is shown in SEQ ID NO.9, and the nucleotide sequence of Hsa_circXPO1_021 is shown in SEQ ID NO.10.

[0027] The experimental method involved in this invention is as follows: 1. Blood sample processing Human peripheral blood was collected in EDTA anticoagulant tubes. Immediately after irradiation, the blood was inoculated into RPMI 1640 medium containing 20% ​​fetal bovine serum and 1% penicillin and streptomycin, with 1.5 mL of whole blood inoculated per 8 mL of medium. The tubes were then incubated at 37°C for the specified time.

[0028] 2. Whole blood total RNA extraction method 1. Total RNA extraction was performed using the RNAprep Pure high-efficiency blood total RNA extraction kit (purchased from Tiangen Biotech (Beijing) Co., Ltd., catalog number DP443). The specific steps are as follows: a. Dilution of erythrocyte lysis buffer: Select an appropriate volume of 10× erythrocyte lysis buffer H according to the volume of the blood sample to be processed (for example, if the volume of the blood sample to be processed is 200 μL, then take 140 μL of 10× erythrocyte lysis buffer H), and dilute it to 1× erythrocyte lysis buffer H with RNase-Free ddH2O.

[0029] b. Remove the human peripheral blood from the incubator and centrifuge at 1000 rpm for 10 min at 4°C. Discard the supernatant; the remaining blood volume in the centrifuge tube should be approximately 1.5 mL. Then add 5 times the volume of 1× erythrocyte lysis buffer H (approximately 7.5 mL) to the centrifuge tube and mix by inverting. Note: For optimal mixing, the volume of the mixture of blood and 1× erythrocyte lysis buffer H should not exceed 3 / 4 of the tube volume. If the white blood cell count is high, the volume of blood used can be reduced proportionally, and the volume of 1× erythrocyte lysis buffer H should be adjusted accordingly.

[0030] c. Incubate on ice for 20 min, vortexing twice during the incubation process.

[0031] d. Centrifuge at 2000 rpm for 10 min at 4℃, remove the supernatant, and the resulting precipitate is a white blood cell precipitate.

[0032] e. Add approximately 3 mL of 1× erythrocyte lysis buffer H to the leukocyte pellet (the volume of 1× erythrocyte lysis buffer H added is twice the volume of whole blood used in step b), resuspend the cells, and perform a second erythrocyte lysis.

[0033] f. Centrifuge at 4℃, 2000 rpm for 10 min to remove the supernatant as completely as possible, otherwise it will affect lysis and subsequent RNA binding to the membrane, resulting in a decrease in the final RNA yield.

[0034] g. Add 600 µL of lysis buffer RLH to the leukocyte pellet (add β-mercaptoethanol to RLH to a final concentration of 1% before the operation, such as adding 10 μL of β-mercaptoethanol to 1 mL of RLH. This lysis buffer is best prepared fresh for use; 600 µL is the volume required for lysing 0.5~1.5 mL of human peripheral blood leukocytes), vortex or mix using a pipette.

[0035] h. Transfer the solution to the CS filter column (place the CS filter column in the collection tube), centrifuge at 12000 rpm for 2 min, discard the CS filter column, and collect the filtrate. Note: To avoid aerosol formation, adjust the pipette to ≥750 µL to ensure that all the solution is transferred to the filter column at once. If there are too many cells, the lysis buffer will become viscous, making it difficult to aspirate.

[0036] i. Add one volume of 70% ethanol to the filtrate (equal to the amount of RLH lysis buffer used in step g), mix well (precipitation may occur at this point), and transfer the resulting solution and precipitate together into the CR4 adsorption column (place the CR4 column in the collection tube). Centrifuge at 12000 rpm for 60 s, discard the waste liquid in the collection tube, and return the CR4 column to the collection tube. Note: When preparing 70% ethanol, please use RNase-free ddH2O. If there is any loss in the volume of the filtrate, please reduce the amount of 70% ethanol accordingly. When transferring the solution and precipitate to the CR4 adsorption column, if the volume is greater than the capacity of the adsorption column, it can be done in two steps.

[0037] j. Add 700 µL of protein removal solution RW1H to the adsorption column CR4 (please check whether ethanol has been added before use), centrifuge at 12000 rpm for 60 s, and discard the waste liquid in the collection tube.

[0038] k. Add 500 µL of wash buffer RW to the adsorption column CR4 (please check that ethanol has been added before use), let stand at room temperature for 2 min, centrifuge at 12000 rpm for 60 s, discard the waste liquid in the collection tube, and put the adsorption column CR4 back into the collection tube. Repeat this step once, for a total of two operations.

[0039] 1. Centrifuge at 12000 rpm for 2 min and discard the waste liquid. Place the CR4 adsorption column at room temperature for several minutes to thoroughly dry any remaining wash solution. Note: After centrifugation, allow the CR4 adsorption column to air dry completely at room temperature. Residual wash solution may affect subsequent experiments such as reverse transcription and quantitative fluorometry.

[0040] m. Transfer the adsorption column CR4 into a new RNase-Free centrifuge tube, add 30 µL of RNase-Free ddH2O, incubate at room temperature for 5 min, and centrifuge at 12000 rpm for 2 min to obtain the RNA solution. Note: The elution buffer volume should not be less than 30 µL; a smaller volume will affect the recovery efficiency. Store the RNA solution at -70℃.

[0041] 3. Reagents and procedures for reverse transcription of cDNA The extracted total RNA was directly subjected to cDNA reverse transcription using a High-Capacity cDNA Reverse Transcription Kit (purchased from Applied Biosystems). TM (Catalog number 4368814). The specific steps are as follows: Prepare the reaction mixture on ice according to Table 1, with a total volume of 20 µL. Perform the reaction in a standard PCR instrument. The reaction program is: 25℃ for 10 min, 37℃ for 120 min, 85℃ for 5 min, cool to 4℃. After the reaction is complete, high-yield cDNA can be obtained and stored at -20℃.

[0042] Table 1 Reverse transcription reaction system

[0043] 4. Detection of circRNA expression levels The expression level of circRNA was detected using real-time quantitative PCR. Primers for different circRNAs are shown in Table 2. The primer design principle in Table 2 is that back-to-back primers across the circRNA reverse splice site can specifically amplify circRNA, avoiding the amplification of the corresponding linear transcript. The reagent used for real-time quantitative PCR was PowerTrack™ SYBR Green Master Mix (purchased from Applied Biosystems). TM (Item number A46109), the reaction system is shown in Table 3. ABI Quant Studio was used. TM 6. The expression levels of different circRNAs were detected using a Flex real-time quantitative PCR instrument. Reaction conditions: 95℃ pre-denaturation for 2 min, followed by 40 cycles of: 95℃ denaturation for 15 s, annealing and extension at 60℃ for a total of 60 s. U6 and GAPDH were used as dual internal controls to normalize the expression levels of each sample. -△△Ct The relative expression levels of circRNA between the irradiation group and the control group were calculated. Three replicates were established for all samples.

[0044] Table 2 Primer Information

[0045] Table 3 Reaction system for real-time quantitative PCR

[0046] Example 1: Screening and validation of circRNA biomarker combinations for radiation dose classification 1. Basic information about the sample This invention collected peripheral blood samples from 30 individuals for the establishment of a circRNA dose classification model. Among them, there were 15 males and 15 females, and their ages were divided into two groups: 20-29 years old and 30-39 years old, with 15 samples in each age group. The median age of the 20-29 year old group was 24 years, and the median age of the 30-39 year old group was 36 years. Informed consent was obtained from all samples collected, which complied with the ethical review requirements of the Institute of Radiation Safety, Chinese Center for Disease Control and Prevention (ethics number: LLSC-NIRP 2024-003).

[0047] 2. Methods of radiation treatment To explore the radiation dose-response and time-response characteristics of candidate circRNA expression levels, radiation doses of 0, 0.5, 1, 2, 4, 6, and 8 Gy were used. 60 Human peripheral blood whole blood samples were uniformly irradiated with Coγ rays in vitro. Total RNA was extracted from the whole blood at 4, 24, 48 and 72 h after irradiation, and the expression level of candidate circRNAs was detected by real-time quantitative PCR.

[0048] 3. Radiation dose response and time response characteristics of candidate circRNA expression levels Radiation treatment and corresponding detection of human peripheral blood samples revealed that the expression level of Hsa_circZDHHC21_004 did not show a significant dose-dependent characteristic 4 h after irradiation, but from 24 h after irradiation, it showed a trend of decreasing with dose, especially at 48 and 72 h after irradiation, where this circRNA showed a significant dose-dependent downregulation. F = 6.811, 8.307, P = 0.000, 0.000;), and the dose-dependent downregulation of circRNA expression levels became more pronounced with increasing post-irradiation time. Figure 1 In the A section, Hsa_circATP5C1_006 and Hsa_circMPP6_020 also exhibited the same dose-response and time-response characteristics as Hsa_circZDHHC21_004, both showing dose-dependent downregulation at 48 and 72 h post-irradiation. F = 2.372, 5.423, P = 0.031, 0.000;F = 2.195, 5.428, P = 0.046, 0.000) Figure 1 (B and C in the original text). This indicates that the expression levels of the above three human peripheral blood circRNAs are affected by ionizing radiation, and their dose-dependent changes appear at 48 and 72 h after irradiation, showing the potential to reflect dose changes. However, the expression of the remaining seven circRNAs did not show significant changes at any of the four post-irradiation time points. Figure 2 This suggests that these circRNAs may not be affected by ionizing radiation and therefore do not have the ability to reflect dose changes.

[0049] 4. Effects of gender and age on the expression levels of Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020 The above experimental results indicate that Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020 are radiation-sensitive circRNAs. Therefore, Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020 were initially identified as candidate circRNAs. However, if they are to be used for dose classification, non-radiation-influencing factors on their expression levels should also be determined. Therefore, the effects of sex and age on the expression levels of these three circRNAs were compared. The results showed that at each dose point and at each post-irradiation time point, there were no statistically significant differences in the expression levels of these three circRNAs between sex and age groups. Figure 3 and Figure 4 This indicates that gender and age do not affect the expression levels of Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020, suggesting that these three circRNAs can be applied to people of different genders and ages when performing dose classification.

[0050] 5. Establish circRNA dose classification models using different machine learning algorithms. Subsequently, a dose classification model was established based on the expression levels of three circRNAs at 48 h and 72 h post-irradiation. The selected machine learning algorithms were Support Vector Machine (SVM) and Random Forest, and four classification models were established. Models 1 and 3 were used to distinguish between unirradiated and irradiated samples at 48 h and 72 h post-irradiation, while models 2 and 4 were used to distinguish between samples with an intensity of around 2 Gy at 48 h and 72 h post-irradiation. The performance of the models built using these two algorithms was then compared. The results showed that, in terms of accuracy, precision, recall, F1 score, and AUC, the four models built using SVM outperformed those built using Random Forest. Therefore, the models built using SVM were selected for double-blind validation (Table 4). Figure 5 and Figure 6 ).

[0051] Table 4 Performance parameters of circRNA dose classification models based on different machine learning algorithms

[0052] 6. Validation of the circRNA dose classification model We validated the generalization ability of the established model using a double-blind method with irradiation doses of 0, 0.7, 3, and 5 Gy. The expression levels of three circRNAs in the model were detected at 48 h and 72 h post-irradiation. Then, models 1 through 4 were used to predict the dose classification of each sample at each time point and dose, and the results were compared with the actual dose classification. The results showed that when distinguishing between unirradiated (G1) and irradiated (G2) samples, models 1 and 3 had higher accuracy in identifying irradiated samples, but their prediction ability for unirradiated samples was poor. This may be related to the imbalance between the number of unirradiated and irradiated samples included in the model training. When distinguishing between samples above and below 2 Gy, model 2 had better accuracy in identifying samples above 2 Gy, while model 4 had better accuracy than model 2. This indicates that the accuracy of the model prediction improves with the extension of post-irradiation time. This may be related to the fact that the dose response of the three circRNA expression levels becomes more pronounced with the extension of post-irradiation time; the stronger the characteristic of the predicted index, the better the model's recognition ability (Tables 5 and 6).

[0053] Table 5. Results of dose classification for unknown samples 48 h post-irradiation using Model 1 and Model 2.

[0054] Table 6. Results of dose classification for unknown samples 72 h post-irradiation using Model 3 and Model 4.

[0055] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A combination of circRNA biomarkers for radiation dose classification, characterized in that, The circRNA biomarker set consists of Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020, the nucleotide sequences of which are shown in SEQ ID NO.1, SEQ ID NO.2, and SEQ ID NO.3, respectively.

2. The use of the circRNA biomarker combination of claim 1 in the preparation of products for radiation dose classification of irradiated persons.

3. The application according to claim 2, characterized in that, The radiation dose classification also requires the collection of peripheral blood samples from the irradiated individuals, which are collected 48-72 hours after exposure.

4. A reagent for detecting the expression level of the circRNA biomarker combination of claim 1.

5. The reagent according to claim 4, characterized in that, The reagents are used to detect the expression levels of each circRNA in the circRNA biomarker combination.

6. The reagent according to claim 5, characterized in that, The reagents include primer pairs specifically for detecting Hsa_circZDHHC21_004, Hsa_circATP5C1_006, and Hsa_circMPP6_020.

7. The use of the reagent according to any one of claims 4-6 in the preparation of a kit for classifying radiation doses of irradiated persons.

8. The application according to claim 7, characterized in that, The radiation dose classification also requires the collection of peripheral blood samples from the irradiated individuals, which are collected 48-72 hours after exposure.

9. A kit for classifying radiation doses of irradiated persons, comprising the reagent of any one of claims 4-6.

10. The reagent kit according to claim 9, characterized in that, The radiation dose classification also requires the collection of peripheral blood samples from the irradiated individuals, which are collected 48-72 hours after exposure.

Images