Methods, apparatus, equipment, storage media and procedures for identifying markers

Abstract

The present application discloses a marker determination method, device, equipment, storage medium and program product. The method comprises: acquiring first expression quantity and second expression quantity of multiple target objects in multiple microRNAs in the perioperative period; determining a first microRNA from the multiple microRNAs according to the first expression quantity, the second expression quantity and a change quantity; inputting the first expression quantity and the second expression quantity into a preset screening model to obtain a second microRNA by screening; for each target object, drawing an expression quantity change trajectory curve of each microRNA of the target object according to the first expression quantity and the second expression quantity; determining a third microRNA according to the expression quantity change trajectory curve; and determining a first candidate target marker according to the first microRNA, the second microRNA and the third microRNA. According to the embodiments of the present application, the long-term clinical events of coronary heart disease are accurately predicted through the change of the determined marker.

Description

Technical Field

[0001] This application belongs to the technical field of clinical medicine, and in particular relates to a method, apparatus, equipment, storage medium and program product for identifying biomarkers. Background Technology

[0002] Coronary heart disease is the leading cause of death among Chinese residents. Its pathogenesis is complex and clinical outcomes vary greatly among individuals. Currently, there is a lack of biomarkers that can accurately assess its risk and prognosis. Large-scale standardized populations, tissue-specific biomarker identification, and accurate detection methods are key to solving this problem.

[0003] For this type of coronary artery disease, coronary artery bypass grafting (CABG) is generally used for treatment. However, the surgery itself is risky, with a high perioperative mortality rate and complications. Moreover, after the surgery, certain adverse events may occur in the long term, affecting the patient's health. Currently, the prediction of the long-term treatment outcome of coronary artery disease mainly relies on medical history information. However, the prediction of the long-term treatment outcome of coronary artery disease based on medical history information is generally not very effective. Therefore, there is an urgent need for a biomarker to accurately predict adverse events after long-term treatment of coronary artery disease by observing changes in the biomarker. Summary of the Invention

[0004] This application provides a biomarker determination method, apparatus, device, storage medium, and program product, which can accurately predict long-term clinical events of coronary heart disease by detecting changes in determined biomarkers.

[0005] On the one hand, embodiments of this application provide a method for determining a marker, the method comprising:

[0006] The first expression level and the second expression level of multiple microRNAs in multiple target subjects during the perioperative period are obtained. The first expression level represents the expression level of multiple microRNAs in the target subjects at a first preset time before surgery, and the second expression level represents the expression level of multiple microRNAs in the target subjects at a second preset time after surgery. The target subjects include patients who have undergone coronary artery bypass grafting.

[0007] Based on the first expression level, the second expression level, and the first change at different preset time points, a first microRNA that differs from the preset control group is identified from a variety of microRNAs. The first change represents the change between the first expression level and the second expression level at different preset time points.

[0008] The first expression level, the second expression level, and the first change level are input into a preset screening model so that the preset screening model can screen for a second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA. The second microRNA represents microRNAs related to cardiovascular and cerebrovascular medical events.

[0009] For each target object, based on the first expression level and the second expression level, plot the trajectory curve of the expression level change of each microRNA in the target object during the perioperative period;

[0010] Based on the expression level change trajectory curve of each target object, a third microRNA is determined, wherein the third microRNA represents a microRNA whose expression level change trajectory differs from that of the preset control group;

[0011] Based on the first microRNA, the second microRNA, and the third microRNA, a first candidate target biomarker is determined to predict the probability of the target subject experiencing adverse cardiovascular and cerebrovascular events within a preset time period.

[0012] Optionally, the step of identifying the first microRNA that differs from the preset control group from a variety of microRNAs based on the first expression level, the second expression level, and the first change at different preset time points includes:

[0013] The first expression level, second expression level, and first change for each microRNA;

[0014] The target subjects were divided into different subgroups according to the type of surgery and the expected event type, and the subgroups included an event group and a control group;

[0015] In each subgroup, the frequency of occurrence of candidate microRNAs in different subgroups was counted. Candidate microRNAs include microRNAs in which the difference between the change or state of each microRNA of the target object in the event group and the change or state of the corresponding microRNA of the target object in the control group is greater than a preset difference.

[0016] Candidate microRNAs with an occurrence frequency greater than a preset frequency are identified as the first microRNA.

[0017] Optionally, the step of inputting the first expression level, the second expression level, and the first change level into a preset screening model, so that the preset screening model screens for the second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA, includes:

[0018] For each microRNA, the first expression level, the second expression level, and the first change level are input into a preset screening model to obtain the correlation coefficient between each microRNA and cardiovascular and cerebrovascular medical events;

[0019] Based on the correlation coefficient and the preset correlation coefficient threshold, multiple second candidate microRNAs are determined;

[0020] The second candidate microRNA that meets the preset expression level screening criteria is identified as the second microRNA. The preset screening criteria include that the correlation coefficient between the second candidate microRNA and cardiovascular and cerebrovascular medical events is greater than a preset threshold.

[0021] Optionally, before inputting the first expression level, the second expression level, and the first change level into a preset screening model, so that the preset screening model screens for the second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA, the method further includes:

[0022] Multiple training samples are obtained, including the state expression level of each microRNA of multiple training subjects during the perioperative period and the corresponding state correlation coefficient at time points. The state correlation coefficient at time points represents the correlation between the state expression level and cardiovascular and cerebrovascular medical events.

[0023] The state expression is input into the initial regression model to obtain the predicted correlation coefficient;

[0024] Based on the predicted correlation coefficient and the historical correlation coefficient, the loss function of the initial regression model is determined;

[0025] If the loss function does not meet the convergence condition, the model parameters of the initial regression model are adjusted to optimize the initial regression model until the loss function meets the convergence condition, thus obtaining the preset screening model.

[0026] Optionally, determining the third microRNA based on the expression level change trajectory curve of each of the target objects includes:

[0027] Obtain the surgery type and cardiovascular medical event type for each target object;

[0028] Multiple target subjects are grouped according to the surgical type and cardiovascular medical event type to obtain multiple second groups;

[0029] For each second group, the expression level change trajectory curve of each microRNA of each target object is compared to obtain the differential expression level change trajectory curve. The differential expression level change trajectory curve represents the expression level change trajectory curve where the change trend of the microRNA is inconsistent with the change trend of the microRNA of the preset control group.

[0030] The microRNA corresponding to the differential expression level change trajectory curve is identified as the third microRNA.

[0031] Optionally, after determining the first candidate biomarker based on the first microRNA, the second microRNA, and the third microRNA, the method further includes:

[0032] Based on the first expression level and the second expression level, a first ratio of any two first candidate biomarkers is determined. The first ratio represents the expression ratio of any two first candidate biomarkers for the same target object within a first preset time or a second preset time.

[0033] Obtain first feature information of the target object, the first feature information including basic information of the target object and treatment method;

[0034] The first feature information and the first ratio are input into a preset risk classification model to perform regression analysis on the first ratio, thereby obtaining the risk probability corresponding to each first feature information and the first ratio with a cardiovascular and cerebrovascular medical event. The preset risk classification model is used to analyze the risk relationship between the first feature information, the first ratio and the cardiovascular and cerebrovascular medical event.

[0035] If the risk probability is not a preset ratio, the first candidate target marker corresponding to the first ratio is determined as the second candidate target marker.

[0036] Optionally, the step of inputting the first feature information and the first ratio into a preset risk classification model to perform regression analysis on the first ratio to obtain the risk probability corresponding to each first feature information and the first ratio and a cardiovascular and cerebrovascular medical event includes:

[0037] For each first feature and first ratio, the regression coefficient of the first ratio in the preset risk classification model is determined using the maximum likelihood estimation method based on the basic information and treatment method.

[0038] Based on the regression coefficients, the risk probability corresponding to each of the first feature information and the first ratio for cardiovascular and cerebrovascular medical events is determined.

[0039] Optionally, after determining the first candidate target marker corresponding to the first ratio as the second candidate target marker, the method further includes:

[0040] Multiple first ratios are combined in any number of ways to obtain multiple verification combinations;

[0041] The first ratio is input into a preset binary classification model to obtain the subject operating characteristic curve corresponding to the first ratio.

[0042] Integrating the working characteristic curve yields the area under the curve corresponding to the first ratio;

[0043] The multiple verification combinations are cross-validated multiple times to obtain the weight corresponding to each first ratio in each verification, and the weight variation coefficient of each verification combination is calculated.

[0044] The second candidate target marker corresponding to the first ratio in the verification combination where the area under the curve is greater than the preset area and the coefficient of variation of the weight is lower than the preset threshold is determined as the final target marker.

[0045] Optionally, after determining the second candidate target marker corresponding to the first ratio in the verification combination where the area under the curve is greater than a preset area and the weighted coefficient of variation is lower than a preset threshold as the final target marker, the method further includes:

[0046] Obtain the third and fourth expression levels of the final target biomarker of the object to be predicted during the perioperative period, as well as the second feature information;

[0047] Based on the second feature information, the third expression level, and the fourth expression level, the risk probability of the subject to be predicted to experience a cardiovascular and cerebrovascular medical event within a preset time period is predicted, and the predicted probability of the cardiovascular and cerebrovascular medical event is obtained.

[0048] Based on the predicted probabilities, a prevention strategy is determined.

[0049] On the other hand, embodiments of this application also provide a reagent kit, the reagent kit comprising:

[0050] MicroRNA reagent, wherein the microRNA reagent is obtained by the marker determination method described in the first aspect.

[0051] Optionally, the microRNAs include miR-425-5p, miR-23a-3p, miR-181a-5p, miR-22-3p, miR-122-5p, miR-99a-5p, miR-499a-5p, and miR-133a-3p.

[0052] On the other hand, embodiments of this application provide a marker determining device, the device comprising:

[0053] The acquisition module is used to acquire the first and second expression levels of multiple microRNAs in multiple target subjects during the perioperative period. The first expression level represents the expression level of multiple microRNAs in the target subject at a first preset time before surgery, and the second expression level represents the expression level of multiple microRNAs in the target subject at a second preset time after surgery. The target subjects include patients who have undergone coronary artery bypass grafting.

[0054] The determination module is used to determine, based on the first expression level, the second expression level, and the first change amount at different preset time points, the first microRNA that differs from the preset control group from a variety of microRNAs, where the first change amount represents the change amount of the first expression level and the second expression level between different preset time points;

[0055] The screening module is used to input the first expression level, the second expression level, and the first change level into a preset screening model, so that the preset screening model can screen for a second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA, wherein the second microRNA represents microRNAs related to cardiovascular and cerebrovascular medical events;

[0056] The plotting module is used to plot the trajectory curve of the expression level change of each microRNA of the target object during the perioperative period based on the first expression level and the second expression level for each target object.

[0057] The determination module is also used to determine a third microRNA based on the expression level change trajectory curve of each target object, wherein the third microRNA represents a microRNA whose expression level change trajectory differs from that of a preset control group;

[0058] The determination module is further configured to determine a first candidate target biomarker based on the first microRNA, the second microRNA, and the third microRNA, so as to predict the probability of the target object experiencing adverse cardiovascular and cerebrovascular events within a preset time period based on the first candidate target biomarker.

[0059] In another aspect, embodiments of this application provide an electronic device, the device comprising: a processor and a memory storing computer program instructions;

[0060] When the processor executes the computer program instructions, it implements the marker determination method as described in the first aspect.

[0061] In another aspect, embodiments of this application provide a computer storage medium storing computer program instructions, which, when executed by a processor, implement the marker determination method as described in the first aspect.

[0062] In another aspect, embodiments of this application provide a computer program product in which instructions, when executed by a processor of an electronic device, cause the electronic device to perform the marker determination method as described in the first aspect.

[0063] The biomarker determination method, apparatus, device, storage medium, and program product of this application embodiment can screen a large number of microRNAs by analyzing the first expression level and second expression level of multiple microRNAs of multiple target objects at different times during the perioperative period, as well as the first change between each preset time point, through differential change, preset screening model, and expression level change trajectory curve, to obtain the first microRNA, second microRNA, and third microRNA. Then, the first candidate target biomarker is determined based on the first microRNA, second microRNA, and third microRNA. This not only reduces the amount of computation, but also accurately predicts the probability of cardiovascular and cerebrovascular adverse events occurring in the target object within a preset time in the future, so as to prevent adverse consequences of cardiovascular and cerebrovascular adverse events from occurring in the target object in advance and ensure the safety of the target object. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart illustrating a marker determination method provided in one embodiment of this application;

[0066] Figure 2 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0067] Figure 3 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0068] Figure 4 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0069] Figure 5 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0070] Figure 6 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0071] Figure 7 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0072] Figure 8 This is a flowchart illustrating a marker determination method provided in another embodiment of this application;

[0073] Figure 9 This is a schematic diagram of the structure of a marker determining device provided in another embodiment of this application;

[0074] Figure 10 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0075] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0076] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0077] To understand the innovation process of this application, a detailed background introduction is provided below.

[0078] Coronary heart disease is the leading cause of death among Chinese residents. Its pathogenesis is complex and clinical outcomes vary greatly among individuals. Currently, there is a lack of biomarkers that can accurately assess its risk and prognosis. Large-scale standardized populations, tissue-specific biomarker identification, and accurate detection methods are key to solving this problem.

[0079] For this type of coronary artery disease, coronary artery bypass grafting (CABG) is generally used for treatment. However, the surgery itself is risky, with a high perioperative mortality rate and complications. Moreover, after the surgery, certain adverse events may occur in the long term, affecting the patient's health. Currently, the prediction of the long-term treatment outcome of coronary artery disease mainly relies on medical history information. However, the prediction of the long-term treatment outcome of coronary artery disease based on medical history information is generally not very effective. Therefore, there is an urgent need for a biomarker to accurately predict adverse events after long-term treatment of coronary artery disease by observing changes in the biomarker.

[0080] Among related technologies, circulating microRNAs (miRNAs) exhibit unique advantages in cardiovascular disease prognostic assessment due to their high stability, tissue specificity, and relevance to pathological processes. For example, miR-1 is positively correlated with the risk of atrial fibrillation after CABG, miR-208a can reflect the degree of myocardial injury, and inflammation-related miRNAs (such as miR-155) can characterize systemic inflammation levels. However, miRNA research in the CABG field still faces the following technical obstacles:

[0081] 1) Insufficient dynamic monitoring: Existing studies mostly use single time points or specific tissue sources for miRNA detection, failing to systematically reveal the full spectrum of dynamic changes in plasma miRNAs during the perioperative period;

[0082] 2) Blind spots in biomarker screening: Key miRNAs associated with long-term adverse events (such as 10-year mortality) and their optimal detection time windows are not yet clear, and there is a lack of long-term follow-up data to support them;

[0083] 3) Bottlenecks in quantitative technology: Traditional qPCR methods suffer from poor reproducibility in detecting low-abundance miRNAs due to differences in reverse transcription efficiency and insufficient stability of internal controls;

[0084] 4) Lack of data standardization: The comparability of cross-platform detection results is low, and there is a lack of effective correction methods for the interference of individual baseline differences (such as age and kidney function) on miRNA expression levels.

[0085] To address the problems of existing technologies, embodiments of this application provide a biomarker determination method, apparatus, device, storage medium, and program product. By analyzing the first and second expression levels of various microRNAs in multiple target subjects at different times during the perioperative period, and using preset differential expression levels, preset screening models, and expression level change trajectory curves, a large number of microRNAs are screened to obtain first, second, and third microRNAs. Then, based on the first, second, and third microRNAs, a first candidate target biomarker is determined. This not only reduces computational load but also allows for accurate prediction of the probability of adverse cardiovascular and cerebrovascular events occurring in the target subject within a preset timeframe, thereby preventing adverse consequences and ensuring the safety of the target subject. The biomarker determination method provided in this application embodiment is described below.

[0086] Figure 1 A flowchart illustrating a marker determination method according to an embodiment of this application is shown. Figure 1 As shown, the marker determination method may include S101-S106:

[0087] S101, obtain the first and second expression levels of multiple microRNAs in multiple target subjects during the perioperative period.

[0088] In this embodiment, the first expression level can represent the expression level of multiple microRNAs in the target subject at a first preset time before surgery, and the second expression level can represent the expression level of multiple microRNAs in the target subject at a second preset time after surgery. The target subject includes patients who have undergone coronary artery disease surgery.

[0089] As an example, the following steps can be followed to determine the construction of the target object queue and sample collection.

[0090] Specifically, when selecting target patients, pre-defined queue criteria can be set. For example, at least three indicators can be set to screen target patients. The three indicators may include: adult patients who receive elective simple coronary artery bypass grafting, patients who sign informed consent forms and are provided with perioperative plasma oxygen pumps, and patients who do not have other cardiac surgeries. Other cardiac surgeries may include valve replacement surgery.

[0091] A large number of patients were screened using the above criteria. After screening, further exclusion criteria were established, which could include: preoperative serum creatinine ≥200 μM, presence of obstructive hepatobiliary disease or hereditary myopathy, and perioperative use of drugs that affect miRNA expression, such as high-dose glucocorticoids. This patient screening process yielded a large number of target subjects. Based on these target subjects, a cohort was constructed, samples were collected and processed, and the samples were sequenced.

[0092] As an example, it is worth noting that human patients possess a large number of microribonucleic acids (miRNAs), and the expression levels of these miRNAs can be obtained from the patient's cardiomyocytes and vascular endothelial cells. In this embodiment, the perioperative period can refer to the time between the patient's decision to undergo surgery and the completion of the surgery and entry into the recovery phase. Specifically, in this embodiment, the first preset time can be any time point within 48 hours before the start of surgery, the second preset time can be 6 hours after surgery, the third preset time is 24 hours after surgery, and the fourth preset time is 48 hours after surgery. In some embodiments, preferably, based on extensive data, the first preset time is 6 hours before surgery, and the second preset time is 6 hours after surgery.

[0093] S102, based on the first expression level, the second expression level, and the first change at different preset time points, identify the first microRNA that differs from the preset control group from a variety of microRNAs;

[0094] In this embodiment, as an example, the first change amount represents the change amount between the first expression amount and the second expression amount at different preset time points;

[0095] In this embodiment, in S102, the target subjects can be divided into an event group and a control group based on whether they experience coronary heart disease-related adverse events post-surgery. Adverse events can be further divided into cardiogenic events and cerebrovascular events. Furthermore, CABG surgery is mainly divided into two methods: on-pump CABG (ONCAB) and off-pump CABG. First, the cohort can be divided into different subgroups using the above grouping method, and the expression levels of each microRNA can be compared to determine whether there are significant differences between the event group and the control group within each subgroup. Second, the expression levels of each microRNA change after coronary heart disease surgery in each target subject. Therefore, by comparing the expression levels between any two time points, different changes can be obtained, thereby determining whether there are significant differences in the changes of each microRNA between the event group and the control group within each subgroup of patients. The frequency of each microRNA in the above-mentioned surgical subgroup, outcome subgroup, time point subgroup, and change subgroup was statistically analyzed. MicroRNAs exceeding a predetermined threshold frequency were identified as the first microRNA, in order to reduce the types of microRNAs unrelated to coronary heart disease surgery and thus reduce the amount of calculation.

[0096] In this embodiment, cardiovascular and cerebrovascular medical events can be some complications after coronary artery disease surgery, including death caused by cardiovascular and cerebrovascular causes, non-fatal myocardial infarction, and non-fatal ischemic stroke.

[0097] S103, the first expression level, the second expression level, and the first change level are input into the preset screening model so that the preset screening model can screen the second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA.

[0098] In some embodiments, in S103, the second microRNA represents a microRNA associated with a cardiovascular or cerebrovascular medical event.

[0099] In some embodiments, in S103, multiple microRNAs can be screened using a preset screening model. In this embodiment, the preset screening model can screen microRNAs based on their correlation with adverse cardiovascular and cerebrovascular events to reduce the computational load of microRNAs and accurately predict adverse cardiovascular and cerebrovascular events using a second microRNA.

[0100] As an example, the preset screening model can be an unconstrained nonlinear optimization model, and the second microRNA represents microRNAs associated with adverse cardiovascular and cerebrovascular events.

[0101] S104. For each target object, plot the trajectory curve of the expression level change of each microRNA during the perioperative period.

[0102] In some embodiments, in S104, each target object will change for each microRNA during the perioperative period as time changes. The expression level change trajectory curve can be fitted by the first expression level, the second expression level, and the first change level. That is, the expression level change trajectory curve can show the changes of each microRNA during the perioperative period.

[0103] S105, based on the expression level change trajectory curve of each target object, determines the third microRNA.

[0104] As an example, the third microRNA represents a microRNA whose expression level change trajectory differs from that of the preset control group.

[0105] In some embodiments, in S105, although there are individual differences among each target object, the expression change trends of different microRNAs in different populations during the perioperative period may be similar. Therefore, the microRNAs with different change trends between the event group and the control group can be identified by the expression change trajectory curve, and the microRNA with different change trends is identified as the third microRNA.

[0106] As an example, non-parametric smoothing algorithms (such as spline regression or Gaussian processes) can be used to fit the non-linear change trajectory of ribonucleic acid (RNA) biomarkers over time. A third microRNA can represent a microRNA whose change trajectory shows a significant difference between the control and event groups; this microRNA can be identified as the third microRNA, and its use has been employed to predict the probability of adverse cardiovascular events.

[0107] S106, Based on the first microRNA, the second microRNA, and the third microRNA, determine the first candidate target biomarker, so as to predict the probability of the target subject experiencing adverse cardiovascular and cerebrovascular events within a preset time period based on the biomarker.

[0108] In some embodiments, in S106, the first candidate target biomarker determined by the first microRNA, the second microRNA, and the third microRNA determined by the above three methods can predict the probability of adverse cardiovascular and cerebrovascular events occurring in the target subject within a future time period. That is, the change of the first candidate target biomarker during the perioperative period can reflect the probability of adverse cardiovascular and cerebrovascular events occurring within a preset time period in the future, thereby improving the accuracy of accurate prediction of adverse cardiovascular and cerebrovascular events.

[0109] In this embodiment, by analyzing the first and second expression levels of various microRNAs of multiple target subjects at different times during the perioperative period, and by directly comparing the state or change of expression levels between groups, using a preset screening model, and analyzing the expression level change trajectory curve, a large number of microRNAs are screened to obtain the first microRNA, the second microRNA, and the third microRNA. Then, based on the first microRNA, the second microRNA, and the third microRNA, a first candidate biomarker is determined. This not only reduces the computational load but also enables the accurate prediction of the probability of adverse cardiovascular and cerebrovascular events occurring in the target subjects within a preset time period using the first candidate biomarker, thereby preventing adverse consequences of adverse cardiovascular and cerebrovascular events in the target subjects in advance and ensuring the safety of the target subjects.

[0110] Reference Figure 2 In other embodiments, in order to screen multiple microRNAs and reduce the computational load while ensuring that the screened microRNAs can more accurately predict adverse cardiovascular and cerebrovascular events, S102 may include:

[0111] S1021, for the first expression level, second expression level and first change level of each microRNA, the target subjects are divided into different subgroups according to the surgical type and expected event type. The subgroups include the event group and the control group.

[0112] S1022, In each subgroup, the frequency of occurrence of candidate microRNAs in different subgroups is counted;

[0113] S1023, the candidate microRNAs with a frequency greater than the preset frequency are identified as the first microRNA.

[0114] In this embodiment, as an example, the candidate microRNAs include microRNAs in which the difference between the change or state of each microRNA of the target object in the event group and the change or state of the corresponding microRNA of the target object in the control group is greater than a preset difference.

[0115] In this embodiment, when screening using preset control changes, the first expression level of each RNA is collected within 48 hours before the target subject undergoes CABG surgery. Then, the second, third, and fourth expression levels of each RNA are collected at 6 hours, 24 hours, and 48 hours after surgery. After that, the changes in the microRNAs of the target patient before and after surgery are determined. Then, the four expression levels (preoperative, 6 hours postoperative, 24 hours postoperative, and 48 hours postoperative) and the six changes (6 hours vs. preoperative, 24 hours vs. preoperative, 48 hours vs. preoperative, 24 hours vs. 6 hours, 48 ​​hours vs. 24 hours, and 48 hours vs. 24 hours) are screened simultaneously.

[0116] Specifically, in this embodiment, in order to ensure the accuracy of screening, patients who have undergone coronary artery disease surgery and have not experienced adverse cardiovascular or cerebrovascular events are identified as the preset control group. Then, the first candidate microRNA is determined by comparing whether there is a significant difference in the expression level and change level. That is, those that are significantly greater than or less than the control group are identified as the first candidate microRNA.

[0117] The frequency of occurrence of the first candidate microRNA was then determined based on the selection frequency of the first candidate microRNA in different subgroup comparisons.

[0118] Specifically, S1022 may include:

[0119] Multiple target subjects were grouped according to surgical type (all patients, ONCAB patients, and OPCAB patients) and coronary artery disease event type (all event types, cardiac events, and cerebral events), resulting in multiple primary groups;

[0120] Obtain the small ribonucleic acid molecules that show significant changes in each first group;

[0121] Based on the summary results of different first groups, the frequency of occurrence of the first candidate microRNA in terms of expression level and change at each preset time point was determined.

[0122] In this embodiment, the overall predictive efficacy of expression levels and changes at different time points is compared to further reduce the detection time window required for subsequent validation, facilitating clinical application. In this example, the predictive efficacy is evaluated by comparing the AUC values ​​of different microRNAs at different time points. The results indicate that preoperative and 6 hours postoperatively are the two time points with the strongest predictive efficacy.

[0123] It is worth noting that the included patients were those who had undergone coronary artery disease surgery five years or more ago. The first microRNA obtained through this screening can predict the probability of adverse cardiovascular and cerebrovascular events occurring at a predetermined time in the future.

[0124] Reference Figure 3 In other embodiments, to improve the accuracy of predicting the probability of adverse cardiovascular events using microRNAs, S103 may include:

[0125] S1031, For each microRNA, the first expression level, the second expression level, and the first change level are input into the preset screening model to obtain the correlation coefficient between each microRNA and cardiovascular and cerebrovascular medical events;

[0126] S1032, Based on the correlation coefficient and the preset correlation coefficient threshold, multiple second candidate microRNAs are determined;

[0127] S1033, the second candidate microRNA that meets the preset expression level screening conditions is identified as the second microRNA.

[0128] In this embodiment, the preset screening criteria include that the correlation coefficient between the second candidate microRNA and cardiovascular and cerebrovascular medical events is greater than a preset threshold.

[0129] In some embodiments, in S1031, the preset screening model can be an unconstrained nonlinear optimization model. By inputting the expression change into the preset screening model, the classification model within the preset screening model can continuously perform classification optimization to obtain the correlation coefficient between each microRNA and adverse cardiovascular events. It is worth noting that the preset screening model is a nonlinear optimization model based on the correlation between the expression change and adverse cardiovascular events.

[0130] Reference Figure 4 Specifically, methods for training a pre-defined screening model may include:

[0131] S401, acquire multiple training samples;

[0132] S402, Input the state expression into the initial regression model to obtain the predicted correlation coefficient;

[0133] S403, Determine the loss function of the initial regression model based on the predicted correlation coefficient and the historical correlation coefficient;

[0134] S404: If the loss function does not meet the convergence condition, adjust the model parameters of the initial regression model to optimize the initial regression model until the loss function meets the convergence condition, thus obtaining the preset screening model.

[0135] In some embodiments, in S401, the training samples may include the state expression levels of each microRNA of multiple training subjects at multiple time points during the perioperative period and the corresponding time-specific correlation coefficients, whereby the correlation coefficients represent the correlation between historical expression levels and cardiovascular and cerebrovascular medical events.

[0136] It is worth noting that the training subjects can be patients who have already undergone coronary artery disease surgery, and these subjects can be patients of all ages, genders, and with all comorbidities, that is, patients involved in all aspects of coronary artery disease surgery.

[0137] In some other embodiments, in S402 and S403, the initial regression model can be a logistic regression model or a support vector machine. There is no limitation here, as long as it can achieve unconstrained nonlinear optimization.

[0138] Specifically, in order to ensure that the initial regression model can classify more accurately, the initial regression model can be initialized, that is, the objective function of the initial regression model can be defined by formula (1):

[0139]

[0140] Where ω is the weight vector of the microRNA, is the loss function value, and m is the number of training samples.

[0141] In other embodiments, the initial regression model can be trained using a regularization algorithm to quickly train the initial regression model and obtain a preset screening model.

[0142] As an example, when ω=0, it means that the training sample is not selected, and when ω>0, it means that the training sample is selected to achieve the screening of microRNAs.

[0143] In some other embodiments, in S404, the loss function can be a cross-entropy loss function or a mean squared error function.

[0144] In other embodiments, the model performance index can be used to determine whether the initial regression model meets the convergence condition. That is, the convergence condition can indicate that the initial training model has been trained or converged. The specific adjustment method and related technology are the same, and will not be elaborated here.

[0145] In some other embodiments, in S1033, after the expression change is input into the preset screening model, the preset screening model can output a correlation coefficient based on the expression change. Then, multiple second candidate microRNAs are determined by the correlation coefficient and the preset correlation coefficient threshold. Specifically, for example, if the correlation coefficient of a certain microRNA is 0.5 and the preset correlation coefficient threshold is 0.3, it means that the microRNA is a second candidate microRNA.

[0146] In addition, to ensure the representativeness of the selected microRNAs, it is necessary to further screen the second candidate microRNAs using preset expression level screening conditions. As an example, the preset expression level screening conditions include that the third expression level of the second candidate microRNA is greater than the preset expression level threshold. For example, if the correlation coefficient of a certain microRNA is 0.5, the third expression level of the microRNA is 500, and the preset expression level threshold is 200, then the microRNA can be identified as the second microRNA.

[0147] Reference Figure 5 In other embodiments, for screening microRNAs, S105 may include:

[0148] S1051, Obtain the surgery type and cardiovascular medical event type for each target object;

[0149] S1052, group multiple target subjects according to the type of surgery and the type of cardiovascular and cerebrovascular event to obtain multiple second groups;

[0150] S1053, For each second group, compare the trajectory curve of the expression level change of each microRNA of each target object to obtain the trajectory curve of the differential expression level change;

[0151] S1054 identified the microRNA corresponding to the differential expression level change trajectory curve as the third microRNA.

[0152] In some embodiments, the implementation of S1051 is the same as that of S1022 described above, and will not be repeated here.

[0153] In some other embodiments, S1052 is implemented in the same way as S1022 described above, and will not be described in detail here.

[0154] In some other embodiments, in S1053, when determining the third microRNA using the expression level change trajectory curve, the expression level change trajectory curve is first plotted for all microRNAs of each target object in each second group, and then the expression level change trajectory curves of different target objects are compared for each microRNA. The differential expression level change curve is determined by the expression level change curves of different target objects for each microRNA.

[0155] For example, if the expression level of a target RNA in the event group first increases and then decreases, but the expression level of the target RNA in the control group first decreases and then increases, then the expression level change trajectory curve can be identified as the differential expression level change trajectory curve.

[0156] In this embodiment, the third microRNA determined by the above method can represent microRNAs with different expression levels among the second groups, and can more accurately predict adverse cardiovascular and cerebrovascular events by identifying changes in microRNAs among different target objects.

[0157] Reference Figure 6In some embodiments, the first, second, and third microRNAs determined by the above method significantly reduce the types of microRNAs. However, many types of the first, second, and third microRNAs still exist. To further reduce computational load, it is necessary to narrow down the candidate range of target biomarkers from the first, second, and third microRNAs to obtain second candidate biomarkers. Simultaneously, for ease of clinical application, the detection platform needs to be replaced with a low-throughput, widely used platform for targeted detection (RT-qPCR). Furthermore, based on the above, the detection time points are further narrowed to two time points: preoperative and 6 hours postoperatively. Specifically, after S106, the method further includes:

[0158] S601, based on the first expression level and the second expression level, determine the first ratio of any two first candidate biomarkers, the first ratio representing the expression level ratio of any two first candidate biomarkers for the same target object within a first preset time or a second preset time.

[0159] S602, Obtain the first characteristic information of the target object, the first characteristic information including the basic information of the target object and the treatment method;

[0160] S603, input the first feature information and the first ratio into the preset risk classification model to perform regression analysis on the first ratio, and obtain the risk probability corresponding to each first feature information and the first ratio and the cardiovascular and cerebrovascular medical event;

[0161] S604, if the risk probability is not a preset ratio, the first candidate target marker corresponding to the first ratio is determined as the second candidate target marker.

[0162] In this embodiment, in S602, the first feature information of the target object is obtained; the first candidate target marker is detected using the RT-qPCR platform in the expanded sample size queue, and the RT-qPCR detection data is standardized using an appropriate method.

[0163] In some embodiments, in S602, the first characteristic information of the target object can be obtained through the hospital's information storage system. The first characteristic information may include the target object's basic information and treatment method. The basic information may include age, gender, comorbidities, and cardiac function, and the treatment method may be a treatment method related to the cause of coronary heart disease surgery.

[0164] In some embodiments, in S602, the RT-qPCR platform for detecting the first candidate biomarker is not unique; commercially available or self-developed dye-based or probe-based RT-qPCR platforms can be used. However, there are differences in detection efficiency, specificity, and sensitivity among the aforementioned different platforms, and the appropriate platform can be selected according to specific needs. In this example, RT-qPCR detection is performed using the miRXES RT-qPCR platform with three specific primers (specific reverse transcription stem-loop primer, specific PCR forward primer, and specific PCR reverse primer). Preferably, this platform can achieve accurate and stable detection of microRNAs with varying abundances.

[0165] In some embodiments, the background ribonucleic acid (RNA) abundance varies among different target subjects, but these differences have little correlation with long-term coronary heart disease events. Therefore, it is necessary to correct for these inherent inter-individual differences. In some instances, this correction is often performed using fixed internal control RNAs widely used in previous studies, such as hsa-miR-16-5p. In this example, our samples showed that the internal control used in previous studies was not applicable because its expression level changed significantly with the surgical time point and was not a stable reference. Therefore, preferably, we used a ratio method to standardize the RT-qPCR detection data. That is, for the 24 first candidate biomarkers, all microRNAs were standardized by pairwise combination and calculating their ratios.

[0166] In this embodiment, 276 preoperative ratios, 6-hour postoperative ratios, and ratio change ranges were obtained for all target objects in the expanded verification internal queue.

[0167] To avoid including too many ratios with low correlation to long-term coronary heart disease, we conducted inter-group comparisons of the three ratios in the event group and the control group, and finally, ratios with a P-value less than 0.05 were included in the next step of evaluation.

[0168] To avoid including highly correlated ratios that would prevent the model from fitting effectively, we performed correlation analysis on the selected ratios and included ratios with a correlation of less than 0.75 in the subsequent analysis.

[0169] Then, the expression levels and changes of the first feature information and the ratios of the first, second and third microRNAs are input into a preset risk classification model. The corresponding risk probabilities are obtained through the preset risk classification model. The preset risk classification model is used to analyze the risk relationship between the first feature information, the first candidate target biomarker and adverse cardiovascular and cerebrovascular events.

[0170] In this embodiment, the preset risk classification model can be the Cox Proportional-Hazards Model (CoxPH model). The proportional risk model is a commonly used statistical model used to analyze the relationship between time-related events and multiple covariates, so as to allow the assessment of the relative impact of primary feature information on risk.

[0171] Specifically, in some embodiments, S603 may include:

[0172] For each first feature and first ratio, the regression coefficient of the first ratio in the preset risk classification model is determined using the maximum likelihood estimation method based on basic information and treatment method.

[0173] Based on the regression coefficients, the risk probability corresponding to each first feature and first ratio for cardiovascular and cerebrovascular medical events is determined.

[0174] In this embodiment, when calculating the risk probability, the first feature information can be used as a covariate, that is, the first feature information is converted into a covariate vector to represent the factors that may affect the risk. Then, the regression coefficient is determined by the maximum likelihood estimation method. After that, the risk probability corresponding to the first feature information and the first candidate target marker and the cardiovascular and cerebrovascular adverse event is determined according to the regression coefficient. The specific calculation formula (2) is as follows:

[0175]

[0176] Where HR is the risk probability and βi is the regression coefficient.

[0177] In some other embodiments, in S603, as an example, a preset ratio of 1 is used. If HR > 1, it indicates that the covariate increases the risk, i.e., increases the probability of adverse cardiovascular and cerebrovascular events. If HR < 1, it indicates that the covariate decreases the risk, i.e., decreases the probability of adverse cardiovascular and cerebrovascular events. If HR = 1, it indicates that the covariate has no effect on the risk, i.e., it is unrelated to the probability of adverse cardiovascular and cerebrovascular events. The comparison between the risk probability and the preset ratio can be used to determine whether the first candidate target biomarker is the ratio of the second candidate biomarker. In this example, 13 miRNA ratios were ultimately screened and confirmed as the second candidate biomarkers (Table 1).

[0178] Table 1. Results of one-way Cox proportional hazards regression analysis of miRNA ratios

[0179]

[0180] By using a pre-set risk model, the first candidate biomarker can be screened again based on the first feature information. The resulting ratio of the second candidate biomarker can not only reduce the amount of computation but also reduce the detection cost for clinical applications, thereby achieving accurate prediction of adverse cardiovascular and cerebrovascular events.

[0181] Reference Figure 7 In other embodiments, to determine the optimal combination of biomarkers for predictive effectiveness and to perform double verification, the method may further include, after S106:

[0182] S701, combine multiple first ratios in any number of ways to obtain multiple verification combinations;

[0183] S702, input the first ratio into the preset binary classification model to obtain the subject operating characteristic curve corresponding to the first ratio;

[0184] S703, Integrate the working characteristic curve to obtain the area under the curve corresponding to the first ratio;

[0185] S704, perform multiple cross-validations on multiple validation combinations to obtain the weight corresponding to each first ratio in each validation, and calculate the weight variation coefficient for each validation combination;

[0186] S705, the second candidate target marker corresponding to the first ratio in the verification combination where the area under the curve is greater than the preset area and the weighted coefficient of variation is lower than the preset threshold is determined as the final target marker.

[0187] In this embodiment, taking 13 groups of expression level ratios as an example, if these 13 groups of expression level ratios are combined in any number of ways, 8191 validation combinations are obtained. Through repeated two-fold cross-validation 10 times, it was determined that when the number of biomarker ratios included in the model is 4-6, the predictive efficacy can reach a high level, and the model stability is also acceptable. Finally, the biomarker combinations with the top 30% of cross-validation AUC and the smallest average coefficient of variation of model weights were selected, namely the combination consisting of 5 miRNA ratios (5-miRatio), including preoperative miR-425-5p / miR-23a-3p and postoperative miR-181a-5p / miR-22-3p, miR-181a-5p / miR-122-5p, miR-99a-5p / miR-122-5p, and miR-499a-5p / miR-133a-3p at 6 hours postoperatively, as the final target biomarkers. In the internal validation queue, the above five expression ratios are used for validation. That is, the five expression ratios are used as a set of input data and input into the preset binary classification model to obtain the receiver operating characteristic curve (ROC curve) corresponding to the expression ratio. Then, by integrating the ROC curve, the area under the curve (AUC value) corresponding to the expression ratio is obtained.

[0188] In some embodiments, the ROC curve is displayed by plotting the True Positive Rate (TPR, also called recall or sensitivity) against the False Positive Rate (FPR, also called 1-specificity) at different thresholds.

[0189] True Rate of Return (TPR) = TP / (TP + FN), which represents the proportion of all positive samples that are correctly predicted as positive.

[0190] The false positive rate (FPR) = FP / (FP + TN) represents the proportion of all negative samples that are incorrectly predicted as positive samples.

[0191] In this context, TP (True Positives) represents true positives, FP (False Positives) represents false positives, FN (False Negatives) represents false negatives, and TN (True Negatives) represents true negatives.

[0192] The AUC value is the area under the ROC curve. The AUC value ranges from 0 to 1, and its meaning is as follows:

[0193] AUC = 1: It indicates that the model can perfectly distinguish positive and negative samples.

[0194] AUC = 0.5: It indicates that the model has no discrimination ability and is equivalent to random guessing.

[0195] AUC < 0.5: It indicates that the model performance is worse than random guessing. This situation is rare in practical applications because the model can be improved by reversing the prediction results.

[0196] 0.5 < AUC < 1: It indicates that the model has discrimination ability. The larger the AUC value, the stronger the discrimination ability of the model.

[0197] In this set of embodiments, AUC has the following advantages:

[0198] (1) Not affected by thresholds: AUC evaluates the performance of the entire model rather than the performance under a specific threshold.

[0199] (2) Robustness: AUC is insensitive to changes in the proportion of positive and negative samples, so it has better robustness on different datasets.

[0200] (3) Intuitive and easy to understand: The AUC value is a numerical value between 0 and 1. The larger the value, the better the model performance, which is easy to understand and compare.

[0201] By using the area under the curve and the weight coefficient to re-screen the final biomarker, it can ensure that the new final biomarker can accurately predict the cardiovascular and cerebrovascular adverse events in the future preset time.

[0202] After internal validation: The AUC of 5-miRatio = 0.663 (95% CI: 0.605–0.721), which is 0.112 higher than NYScore (P = 0.006). External validation: AUC = 0.684 (95% CI: 0.564–0.805), and the NRI = 0.213 (P = 0.042) after combination with NYScore.

[0203] Refer to Figure 8 , in some other embodiments, after S106, the method may further include:

[0204] S801, obtaining the third expression level and the fourth expression level of the final biomarker of the object to be predicted during the perioperative period and the second feature information;

[0205] S802, predicting the probability of the object to be predicted having cardiovascular and cerebrovascular adverse events within the future preset time period according to the second feature information, the third expression level and the fourth expression level, and obtaining the prediction probability of the cardiovascular and cerebrovascular adverse events;

[0206] S803 determines prevention strategies based on predicted probabilities.

[0207] In this embodiment, after the final biomarker is determined in the above manner, the probability of adverse cardiovascular and cerebrovascular events that may occur in the subject to be predicted within a preset time period in the future (such as the perioperative long-term period of five years after surgery) can be obtained through the final biomarker. If the predicted probability is greater than the preset threshold, it means that the subject to be predicted will have adverse cardiovascular and cerebrovascular events in the preset time period in the future. Preventive strategies can be made in advance for adverse cardiovascular and cerebrovascular events that may occur in the future, thereby ensuring the life safety of the subject to be predicted.

[0208] In this application embodiment, a kit is also provided, the kit including microRNA reagent, wherein the microRNA reagent is obtained by the above-described marker determination method.

[0209] Preferably, the microRNAs may include miR-425-5p, miR-23a-3p, miR-181a-5p, miR-22-3p, miR-122-5p, miR-99a-5p, miR-499a-5p, and miR-133a-3p.

[0210] In the embodiments of this application, the steps for making the reagent kit are similar to those of related technologies, except for the microRNA, so the process of making the reagent kit will not be described in detail here.

[0211] Based on the marker determination method provided in the above embodiments, this application also provides a specific implementation of the marker determination device 900. Please refer to the following embodiments.

[0212] Reference Figure 9 The marker determining device 900 provided in this application embodiment may include:

[0213] The acquisition module 901 is used to acquire the first expression level and the second expression level of multiple microRNAs in multiple target subjects during the perioperative period. The first expression level represents the expression level of multiple microRNAs in the target subjects at a first preset time before surgery, and the second expression level represents the expression level of multiple microRNAs in the target subjects at a second preset time after surgery. The target subjects include patients who have undergone coronary artery bypass grafting.

[0214] The determination module 902 is used to determine the first microRNA that differs from the preset control group from a variety of microRNAs based on the first expression level, the second expression level, and the first change amount at different preset time points. The first change amount represents the change amount of the first expression level and the second expression level between different preset time points.

[0215] The screening module 903 is used to input the first expression level, the second expression level, and the first change level into the preset screening model, so that the preset screening model can screen the second microRNA based on the first expression level, the second expression level, and the first change level of each microRNA. The second microRNA represents microRNAs related to cardiovascular and cerebrovascular medical events.

[0216] The plotting module 904 is used to plot the trajectory curve of the expression level change of each microRNA of the target object during the perioperative period based on the first expression level and the second expression level.

[0217] The determination module 902 is also used to determine the third microRNA based on the expression level change trajectory curve of each target object. The third microRNA represents the microRNA whose expression level change trajectory differs from the preset control group.

[0218] The determination module 902 is further configured to determine a first candidate target biomarker based on the first microRNA, the second microRNA, and the third microRNA, so as to predict the probability of a target subject experiencing adverse cardiovascular and cerebrovascular events within a preset time period based on the first candidate target biomarker.

[0219] As an optional implementation, the determining module 902 can also be used for:

[0220] The first expression level, second expression level, and first change level for each microRNA;

[0221] The target subjects were divided into different subgroups according to the type of surgery and the type of expected event. The subgroups included an event group and a control group.

[0222] In each subgroup, the frequency of occurrence of candidate microRNAs in different subgroups was counted. Candidate microRNAs include microRNAs in which the difference between the change or state of each microRNA of the target object in the event group and the change or state of the corresponding microRNA of the target object in the control group is greater than a preset difference.

[0223] Candidate microRNAs with a frequency greater than a preset frequency are identified as the first microRNA.

[0224] As an optional implementation, the determining module 902 can also be used for:

[0225] For each microRNA, the first expression level, the second expression level, and the first change level are input into a preset screening model to obtain the correlation coefficient between each microRNA and cardiovascular and cerebrovascular medical events.

[0226] Based on the correlation coefficient and a preset correlation coefficient threshold, multiple second candidate microRNAs were identified.

[0227] The second candidate microRNA that meets the preset expression level screening criteria is identified as the second microRNA. The preset screening criteria include that the correlation coefficient between the second candidate microRNA and cardiovascular and cerebrovascular medical events is greater than a preset threshold.

[0228] As an optional implementation, the determining module 902 can also be used for:

[0229] Multiple training samples were obtained, including the historical expression levels of each microRNA of multiple training subjects during the perioperative period and the corresponding historical correlation coefficients. The historical correlation coefficients represent the correlation between historical expression levels and preset medical events.

[0230] Historical expression levels are input into the initial regression model to obtain the predicted correlation coefficient;

[0231] The loss function of the initial regression model is determined based on the predicted correlation coefficient and the historical correlation coefficient.

[0232] If the loss function does not meet the convergence condition, adjust the model parameters of the initial regression model to optimize the initial regression model until the loss function meets the convergence condition, thus obtaining the preset screening model.

[0233] As an optional implementation, the determining module 902 can also be used for:

[0234] Obtain the surgery type and cardiovascular medical event type for each target object;

[0235] Multiple target subjects were grouped according to the type of surgery and the type of cardiovascular and cerebrovascular medical event, resulting in multiple secondary groups;

[0236] For each second group, the expression level change trajectory curves of each microRNA in each target object are compared to obtain the differential expression level change trajectory curves. The differential expression level change trajectory curves represent the expression level change trajectory curves where the change trend of microRNA is inconsistent with the change trend of microRNA in the preset control group.

[0237] The microRNA corresponding to the differential expression level change trajectory curve was identified as the third microRNA.

[0238] As an optional implementation, the determining module 902 can also be used for:

[0239] Based on the first expression level and the second expression level, a first ratio of any two first candidate biomarkers is determined. The first ratio represents the expression ratio of any two first candidate biomarkers for the same target object within a first preset time or a second preset time.

[0240] Obtain the first characteristic information of the target object, which includes the basic information of the target object and the treatment method;

[0241] The first feature information and the first ratio are input into the preset risk classification model to perform regression analysis on the first ratio, so as to obtain the risk probability corresponding to each first feature information and the first ratio and the cardiovascular and cerebrovascular medical event. The preset risk classification model is used to analyze the risk relationship between the first feature information, the first ratio and the cardiovascular and cerebrovascular medical event.

[0242] If the risk probability is not a preset ratio, the first candidate target marker corresponding to the first ratio is determined as the second candidate target marker.

[0243] As an optional implementation, the determining module 902 can also be used for:

[0244] For each first feature and first ratio, the regression coefficient of the first ratio in the preset risk classification model is determined using the maximum likelihood estimation method based on basic information and treatment method.

[0245] Based on the regression coefficients, the risk probability corresponding to each first feature and first ratio for cardiovascular and cerebrovascular medical events is determined.

[0246] As an optional implementation, the determining module 902 can also be used for:

[0247] Multiple first ratios are combined in any number of ways to obtain multiple verification combinations;

[0248] The first ratio is input into the preset binary classification model to obtain the subject operating characteristic curve corresponding to the first ratio;

[0249] Integrate the working characteristic curve to obtain the area under the curve corresponding to the first ratio;

[0250] Multiple validation combinations are cross-validated multiple times to obtain the weight corresponding to each first ratio in each validation, and the coefficient of variation of the weights for each validation combination is calculated.

[0251] The second candidate target marker corresponding to the first ratio in the verification combination where the area under the curve is greater than the preset area and the coefficient of variation of the weight is lower than the preset threshold is determined as the final target marker.

[0252] As an optional implementation, the determining module 902 can also be used for:

[0253] Obtain the third and fourth expression levels of the final target biomarker of the subject to be predicted during the perioperative period, as well as the second feature information;

[0254] Based on the second feature information, the third expression level, and the fourth expression level, the risk probability of cardiovascular and cerebrovascular medical events occurring in the subject within a preset time period is predicted, and the predicted probability of cardiovascular and cerebrovascular medical events is obtained.

[0255] Based on the predicted probability, prevention strategies are determined.

[0256] Figure 10 A schematic diagram of the hardware structure of the electronic device provided in an embodiment of this application is shown.

[0257] An electronic device may include a processor 1001 and a memory 1002 storing computer program instructions.

[0258] Specifically, the processor 1001 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0259] Memory 1002 may include mass storage for data or instructions. For example, and not limitingly, memory 1002 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, memory 1002 may include removable or non-removable (or fixed) media, or memory 1002 may be non-volatile solid-state memory. Memory 1002 may be internal or external to the integrated gateway disaster recovery device.

[0260] In one instance, memory 1002 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0261] Memory 1002 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the marker determination method according to the first aspect of this disclosure.

[0262] The processor 1001 reads and executes computer program instructions stored in the memory 1002 to achieve... Figure 1 A marker determination method in the illustrated embodiment.

[0263] In one example, the electronic device may also include a communication interface 1003 and a bus 1004. For example, Figure 10 As shown, the processor 1001, memory 1002, and communication interface 1003 are connected through bus 1004 and complete communication with each other.

[0264] The communication interface 1003 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0265] Bus 1004 includes hardware, software, or both, that couples components of an electronic device together. For example, and not as a limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 1004 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0266] The electronic device can execute the marker determination method in the embodiments of this application, thereby achieving the combination Figures 1-9 The described method and apparatus for determining markers.

[0267] Furthermore, in conjunction with the marker determination methods in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the marker determination methods in the above embodiments.

[0268] In an optional embodiment, in conjunction with the marker determination method in the above embodiments, this application embodiment can provide a computer program product to implement it. The instructions in the computer program product are executed by the processor of an electronic device, enabling the electronic device to implement any of the marker determination methods in the above embodiments.

[0269] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0270] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0271] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0272] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0273] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A marker determination method for non-diagnostic therapeutic purposes, characterized by, The method comprises the following steps: obtaining a plurality of target objects in the perioperative period of a plurality of microRNAs The first expression amount and the second expression amount, the first expression amount represents the expression amount of a plurality of microRNAs in the target object at the first preset time before the operation, and the second expression amount represents the expression amount of a plurality of microRNAs in the target object at the second preset time after the operation, and the target object includes a patient treated by coronary artery bypass grafting; According to the first expression amount, the second expression amount and the first change amount at different preset time points, determine the first microRNA which is different from the preset control group from a plurality of microRNAs, and the first change amount represents the change amount between the first expression amount and the second expression amount at different preset time points; Input the first expression amount, the second expression amount and the first change amount into the preset screening model, so that the preset screening model screens the second microRNA according to the first expression amount, the second expression amount and the first change amount of each microRNA, and the second microRNA represents the microRNA related to cardiovascular and cerebrovascular medical events; For each target object, draw the expression amount change trajectory curve of each microRNA in the target object in the perioperative period according to the first expression amount and the second expression amount; According to the expression amount change trajectory curve of each target object, determine the third microRNA, which represents the microRNA with different expression amount change trajectory from the preset control group; According to the first microRNA, the second microRNA and the third microRNA, determine the first candidate target marker, and predict the probability of the target object occurring cardiovascular adverse events within a future preset time according to the first candidate target marker.

2. The method of claim 1, wherein, According to the first expression amount, the second expression amount and the first change amount at different preset time points, determine the first microRNA which is different from the preset control group from a plurality of microRNAs, and the first change amount represents the change amount between the first expression amount and the second expression amount at different preset time points; For the first expression amount, the second expression amount and the first change amount of each microRNA, the target object is divided into different subgroups according to the type of operation and the type of expected event, and the subgroup includes the event group and the control group; In each subgroup, the occurrence frequency of the candidate microRNA in different subgroups is counted, wherein the candidate microRNA includes the microRNA whose change amount or state amount of each microRNA in the target object in the event group is greater than the preset difference value from the change amount or state amount of the corresponding microRNA in the target object in the control group; The candidate microRNA with an occurrence frequency greater than a preset frequency is determined as the first microRNA.

3. The method of claim 1, wherein, The first expression amount, the second expression amount and the first change amount are input into the preset screening model, so that the preset screening model screens the second microRNA according to the first expression amount, the second expression amount and the first change amount of each microRNA. inputting the first expression amount, the second expression amount and the first change amount into a preset screening model to obtain a correlation coefficient of each micro ribonucleic acid and the cardiovascular medical event; determining a plurality of second candidate micro ribonucleic acids according to the correlation coefficient and a preset correlation coefficient threshold; determining a second micro ribonucleic acid from the second candidate micro ribonucleic acid that meets a preset expression amount screening condition, wherein the preset expression amount screening condition comprises that the correlation coefficient of the second candidate micro ribonucleic acid and the cardiovascular medical event is greater than a preset threshold.

4. The method of claim 3, wherein, Before the first expression amount, the second expression amount and the first change amount are input into the preset screening model to enable the preset screening model to screen the second micro ribonucleic acid according to the first expression amount, the second expression amount and the first change amount of each micro ribonucleic acid, the method further comprises: obtaining a plurality of training samples, wherein the training samples comprise a state expression amount of each micro ribonucleic acid of a plurality of training objects at a plurality of time points during the perioperative period and a corresponding time point-specific correlation coefficient, and the correlation coefficient represents the correlation between the state expression amount and the cardiovascular medical event; inputting the state expression amount into an initial regression model to obtain a predicted correlation coefficient; determining a loss function of the initial regression model according to the predicted correlation coefficient and a historical correlation coefficient; in a case where the loss function does not meet a convergence condition, adjusting a model parameter of the initial regression model to optimize the initial regression model until the loss function meets the convergence condition, thereby obtaining a preset screening model.

5. The method according to any one of claims 1 to 4, characterized in that, The determining of the third micro ribonucleic acid according to the expression amount change trajectory curve of each target object comprises: obtaining a surgery type and a cardiovascular medical event type of each target object; grouping a plurality of target objects according to the surgery type and the cardiovascular medical event type to obtain a plurality of second groups; for each second group, comparing the expression amount change trajectory curve of each micro ribonucleic acid of each target object to obtain a differential expression amount change trajectory curve, wherein the differential expression amount change trajectory curve represents an expression amount change trajectory curve in which the change trend of the micro ribonucleic acid is inconsistent with the change trend of a micro ribonucleic acid of a preset control group; determining a third micro ribonucleic acid from the micro ribonucleic acid corresponding to the differential expression amount change trajectory curve.

6. The method of claim 5, wherein, After the determining of the first candidate target marker according to the first micro ribonucleic acid, the second micro ribonucleic acid and the third micro ribonucleic acid, the method further comprises: determining a first ratio of any two first candidate target markers based on the first expression amount and the second expression amount, wherein the first ratio represents the expression amount ratio of any two first candidate markers within a first preset time or a second preset time for the same target object; obtaining first feature information of the target object, wherein the first feature information comprises basic information and a treatment method of the target object; inputting the first characteristic information and the first ratio into a preset risk classification model to perform regression analysis on the first ratio, to obtain a risk probability corresponding to each of the first characteristic information and the first ratio and a cardiovascular medical event, wherein the preset risk classification model is used to analyze a risk relationship among the first characteristic information, the first ratio, and the cardiovascular medical event; in a case where the risk probability is not a preset ratio, determining the first candidate target marker corresponding to the first ratio as a second candidate target marker.

7. The method of claim 6, wherein, The method further comprises: determining, for each of the first characteristic information and the first ratio, a regression coefficient of the first ratio in the preset risk classification model based on the basic information and the treatment mode by using a maximum likelihood estimation method; based on the regression coefficient, determining a risk probability corresponding to each of the first characteristic information and the first ratio and the cardiovascular medical event.

8. The method of claim 6, wherein, After determining the first candidate target marker corresponding to the first ratio as the second candidate target marker, the method further comprises: combining the plurality of first ratios in any number of combinations to obtain a plurality of verification combinations; inputting the first ratio into a preset binary classification model to obtain a receiver operating characteristic curve corresponding to the first ratio; integrating the operating characteristic curve to obtain an area under the curve corresponding to the first ratio; performing multiple cross-validations on the plurality of verification combinations to obtain a weight corresponding to each of the first ratios in each validation, and calculate a weight coefficient of variation of each verification combination; determining, as a final target marker, a second candidate target marker corresponding to a first ratio in a verification combination in which the area under the curve is greater than a preset area and the weight coefficient of variation is lower than a preset threshold.

9. The method of claim 8, wherein, After determining, as the final target marker, the second candidate target marker corresponding to the first ratio in the verification combination in which the area under the curve is greater than the preset area and the weight coefficient of variation is lower than the preset threshold, the method further comprises: obtaining a third expression level and a fourth expression level of the final target marker of a to-be-predicted object in a perioperative period and second characteristic information; predicting, according to the second characteristic information, the third expression level, and the fourth expression level, a risk probability of the to-be-predicted object of having a cardiovascular medical event in a future preset time period to obtain a prediction probability of the cardiovascular medical event; determining a prevention strategy based on the prediction probability.

10. A kit characterized in that, The kit comprises a micro ribonucleic acid reagent, wherein the micro ribonucleic acid reagent is obtained by the marker determination method according to any one of claims 1-9; and the micro ribonucleic acid comprises miR-425-5p, miR-23a-3p, miR-181a-5p, miR-22-3p, miR-122-5p, miR-99a-5p, miR-499a-5p, and miR-133a-3p.

11. A marker determination apparatus characterized by comprising: The device comprises: an acquisition module configured to acquire first expression amounts and second expression amounts of a plurality of microRNAs of a plurality of target objects in a perioperative period, the first expression amounts representing expression amounts of the microRNAs at a first preset time before surgery of the target objects, and the second expression amounts representing expression amounts of the microRNAs at a second preset time after surgery of the target objects, the target objects including patients treated by coronary artery bypass grafting; a determination module configured to determine, from the microRNAs, first microRNAs different from a preset control group according to the first expression amounts, the second expression amounts, and first change amounts between different preset time points, the first change amounts representing change amounts between the first expression amounts and the second expression amounts at the different preset time points; a screening module configured to input the first expression amounts, the second expression amounts, and the first change amounts into a preset screening model, so that the preset screening model screens second microRNAs according to the first expression amounts, the second expression amounts, and the first change amounts of each microRNA, the second microRNAs representing microRNAs related to cardiovascular and cerebrovascular medical events; a drawing module configured to draw, for each of the target objects, an expression amount change trajectory curve of each microRNA of the target object in the perioperative period according to the first expression amounts and the second expression amounts; the determination module is further configured to determine third microRNAs according to the expression amount change trajectory curves of each of the target objects, the third microRNAs representing microRNAs with different expression amount change trajectories from the preset control group; the determination module is further configured to determine first candidate target markers according to the first microRNAs, the second microRNAs, and the third microRNAs, so as to predict probabilities of the target objects suffering from cardiovascular and cerebrovascular adverse events in a future preset time according to the first candidate target markers.

12. An electronic device, comprising: The device comprises a processor and a memory storing computer program instructions; The processor executes the computer program instructions to implement the marker determination method according to any one of claims 1-9.

13. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer program instructions, and the computer program instructions are executed by a processor to implement the marker determination method according to any one of claims 1-9.

14. A computer program product, characterised in that, The instructions in the computer program product are executed by a processor of an electronic device to enable the electronic device to perform the marker determination method according to any one of claims 1-9.

Images