The use of metabolic markers in predicting the clinical stage of chronic kidney disease

Abstract

The present invention belongs to the field of biomedicine technology, and discloses the use of metabolic markers in preparing products for detecting the clinical staging of chronic kidney disease. A marker for predicting the clinical staging of chronic kidney disease, wherein the marker is selected from one or more of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane. The present invention uses volatile organic compounds in exhaled breath as markers to predict the clinical staging of chronic kidney disease, and has the advantages of convenient multiple collection and easier pre-processing, and therefore has important clinical application value.

Description

Technical Field

[0001] The present invention belongs to the field of biomedicine technology, and specifically relates to the use of metabolic markers in predicting the clinical stage of chronic kidney disease. Background Art

[0002] The incidence of chronic kidney disease is increasing worldwide and has become a major global public health problem.

[0003] At present, the commonly used diagnostic methods for renal function and renal injury in clinical practice are still mainly laboratory tests, imaging tests and renal tissue biopsy. Among them, blood tests, urine tests, eGFR (estimated glomerular filtration rate) and renal tissue biopsy are the most commonly used methods. eGFR mainly reflects the function of glomeruli to clear metabolites, but Scr (serum creatinine) is affected by many factors, and it only starts to increase when the glomerular filtration function decreases to 50% of the normal level, and it cannot reflect the early stage of renal injury; and renal tissue biopsy, as an invasive examination, is not easily accepted by patients without any clinical discomfort symptoms, and when performing renal tissue biopsy, different doctors take different parts of the sample, which may lead to different results, affecting the clinician's diagnosis of the disease and the choice of treatment options. The collection of blood and urine is inconvenient, and the pre-treatment process is complicated, making it difficult to perform high-frequency tests. In addition, the markers in blood and urine have problems with insufficient sensitivity and specificity. The existence of these problems urgently requires people to find more accurate, more practical and more convenient biomarkers that can evaluate renal function.

[0004] As an emerging discipline, metabolomics has gradually become a hot topic of research at home and abroad. It mainly studies the body's metabolites and corresponding metabolic pathways. By qualitatively or quantitatively analyzing the changes in small molecule metabolites in tissues or body fluids, it can discover the abnormal metabolic pathways of the body that are most relevant to the changes, gain a deeper understanding of the pathophysiological mechanisms in the development of diseases, and look for disease-specific biomarkers. Commonly used techniques for metabolomics analysis of metabolites include nuclear magnetic resonance (NMR) technology, mass spectrometry (MS) technology, chromatography-mass spectrometry technology, and the combination of multiple analysis platforms. By analyzing the spectra of a series of samples and combining pattern recognition to determine the pathophysiological state of an organism, it is possible to find related biomarkers and provide an early warning signal for related diseases.

[0005] Chronic kidney disease is divided into five stages according to the glomerular filtration rate: stage 1, stage 2, stage 3, stage 4 and stage 5, with increasing severity. Therefore, how to accurately and effectively conduct clinical staging to guide clinical staging as well as clinical diagnosis and treatment has become an urgent problem that needs to be solved. Summary of the invention

[0006] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a use of metabolic markers in predicting the clinical stage of chronic kidney disease, in order to solve the problems existing in the prior art.

[0007] To achieve the above purpose, the present invention specifically adopts the following technical solutions.

[0008] One of the purposes of the present invention is to protect a marker for predicting the clinical stage of chronic kidney disease, wherein the marker is selected from one or more of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane.

[0009] In certain embodiments of the present invention, the clinical stages are stage 1, stage 3, and stage 5.

[0010] In certain embodiments of the present invention, the marker is derived from the exhaled breath of a chronic kidney disease patient, preferably, the alveolar exhaled breath.

[0011] The second object of the present invention is to protect the use of substances for detecting markers as described above in the preparation of products for predicting the clinical stage of chronic kidney disease.

[0012] In certain embodiments of the present invention, the substance for detecting a marker includes a substance for detecting the volume or mass concentration of the marker.

[0013] In certain embodiments of the present invention, the clinical stages are stage 1, stage 3, and stage 5.

[0014] The third object of the present invention is to protect a product for predicting the clinical stage of chronic kidney disease, wherein the product comprises a substance for detecting a marker for the use as described above.

[0015] In certain embodiments of the present invention, the product comprises at least one of a kit, a membrane strip, a chip and a detection system.

[0016] Preferably, the detection system is selected from a gas sensing system or a gas chromatograph.

[0017] In certain embodiments of the present invention, the clinical stages are stage 1, stage 3, and stage 5.

[0018] A fourth object of the present invention is to protect the use of the markers as described above or the substances for detecting the markers as described above in screening drugs for treating and / or preventing chronic kidney disease.

[0019] The fifth object of the present invention is to protect a method for screening the markers described above, wherein the method detects the organic matter content in samples of patients with clinical stage 1, stage 3 and stage 5 chronic kidney disease, performs differential analysis on the organic matter content, and obtains differential organic matter, which is the marker.

[0020] In certain embodiments of the invention, the sample is exhaled breath.

[0021] In certain embodiments of the present invention, the detection method is gas chromatography-mass spectrometry, and the detection method comprises the following steps:

[0022] 1-1) Enriching exhaled air using adsorbent;

[0023] 1-2) A thermal desorber is used to perform thermal desorption on the adsorbent after enriching the exhaled breath, and then a gas chromatography-mass spectrometer is used for detection. The detection results are qualitatively and quantitatively analyzed, and the types of organic matter are obtained according to the qualitative analysis, and the content of organic matter is obtained according to the quantitative analysis.

[0024] The sixth object of the present invention is to protect a method for constructing a chronic kidney disease clinical staging prediction model, the method comprising:

[0025] 1) constructing a sample data set based on the detection amount of the markers as described above in the subject samples, wherein the subjects include patients with confirmed clinical stage 1, stage 3, and stage 5 chronic kidney disease;

[0026] 2) randomly dividing the sample data set into a test set and a training set, using a machine learning method to learn the training set to obtain the chronic kidney disease clinical staging prediction model.

[0027] In certain embodiments of the present invention, the sample is selected from exhaled air, preferably alveolar exhaled air.

[0028] In certain embodiments of the present invention, the machine learning method includes at least one of a support vector machine algorithm, a nearest neighbor algorithm, and a linear discriminant analysis algorithm.

[0029] In certain embodiments of the present invention, the step 3) is also included in which the chronic kidney disease clinical staging prediction model is tested using a test set.

[0030] The seventh object of the present invention is to protect a method for predicting the clinical stage of chronic kidney disease, comprising the following steps:

[0031] S1. Obtaining the detection amount data of the markers as described above in the sample of the subject;

[0032] S2. Applying the chronic kidney disease clinical staging prediction model obtained by the construction method described above to process the detection data to output the chronic kidney disease clinical staging result;

[0033] In certain embodiments of the present invention, the sample is exhaled air, preferably alveolar exhaled air.

[0034] The eighth object of the present invention is to protect a system for predicting the clinical stage of chronic kidney disease, the system comprising:

[0035] A data module, used to obtain the detection amount data of the markers as described above in the sample of the subject;

[0036] The evaluation module processes the detection data using the chronic kidney disease clinical staging prediction model obtained by the construction method described above to output the chronic kidney disease clinical staging results.

[0037] The ninth object of the present invention is to protect a device, comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the device performs the prediction method described above or the construction method described above.

[0038] A tenth object of the present invention is to protect a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the prediction method as described above or the construction method as described above.

[0039] In the present invention, a large amount of organic matter in the exhaled breath of patients with clinical stage 1, clinical stage 3 and clinical stage 5 of chronic kidney disease was detected, and it was found that the contents of seven volatile organic compounds, namely n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, were significantly different between stages 1, 3 and 5 of chronic kidney disease. Therefore, these seven volatile organic compounds in the exhaled breath were used as markers to predict clinical stages 1, 3 and 5 of chronic kidney disease. Furthermore, a prediction model for the clinical staging of chronic kidney disease was established based on the contents of the seven volatile organic compounds, which can quickly, efficiently and objectively monitor the progress of the chronic kidney disease population in stages 1, 3 and 5. Compared with blood and urine tests, the present invention uses volatile organic compounds in exhaled breath as markers to predict the clinical staging of chronic kidney disease, which has the advantages of convenient multiple collection and easier pre-processing. The present invention has important clinical application value.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1) The markers for predicting the clinical stage of chronic kidney disease of the present invention can be applied to large-scale initial screening of stage 1, stage 3 and stage 5 chronic kidney disease.

[0042] 2) Compared with blood and urine analysis, the present invention predicts the clinical analysis of chronic kidney disease by detecting exhaled breath markers. The sampling and analysis of markers has the advantages of convenience and immediacy, and also greatly improves the compliance and comfort of the subjects; the collection process is non-invasive, simple and fast.

[0043] 3) The chronic kidney disease clinical staging prediction model of the present invention has an accuracy rate of up to 76.3%, and the harmonic means of precision and recall are both higher than 50%, and can specifically identify stage 1, stage 3 and stage 5 patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 Shown is a diagram of the exhaled air collection device of the present invention.

[0045] Figure 2 Shown is the chromatogram of exhaled breath in Example 1 of the present invention. DETAILED DESCRIPTION

[0046] The following is a description of the implementation of the present invention by means of specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0047] Before further describing the specific embodiments of the present invention, it should be understood that the scope of protection of the present invention is not limited to the specific embodiments described below; it should also be understood that the terms used in the examples of the present invention are intended to describe specific embodiments, rather than to limit the scope of protection of the present invention. The test methods in the following examples without specifying specific conditions are generally carried out under conventional conditions or under conditions recommended by the manufacturers.

[0048] When the embodiments give numerical ranges, it should be understood that, unless otherwise specified in the present invention, both endpoints of each numerical range and any numerical value between the two endpoints can be selected. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those generally understood by those skilled in the art. In addition to the specific methods, equipment, and materials used in the embodiments, according to the grasp of the prior art by those skilled in the art and the record of the present invention, any methods, equipment, and materials of the prior art similar or equivalent to the methods, equipment, and materials described in the embodiments of the present invention can also be used to realize the present invention.

[0049] The underlying principle of the present invention is that after a human tissue becomes diseased, it will cause changes in the body's metabolism, resulting in changes in the concentration of some volatile organic compounds in the blood. The volatile substances can pass through the blood-air barrier in the lungs, so their concentration in the exhaled air will change. Therefore, the disease can be identified by detecting changes in the concentration of certain volatile organic compounds in the exhaled air.

[0050] Based on the current need for non-invasive detection of clinical stages of chronic kidney disease mentioned in the background technology section, the present invention detects a large amount of organic matter in the exhaled breath of patients with clinical stage 1, clinical stage 3 and clinical stage 5 of chronic kidney disease, and finds that there are significant differences in the contents of seven volatile organic compounds, namely n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, between stages 1, 3 and 5 of chronic kidney disease. Therefore, using these seven volatile organic compounds in exhaled breath as markers for predicting clinical stages 1, 3 and 5 of chronic kidney disease has important clinical application value. Furthermore, a prediction model for the clinical stage of chronic kidney disease was established based on the contents of the seven volatile organic compounds. The accuracy of the prediction model is relatively high, and it is therefore suitable for quickly, efficiently and objectively distinguishing clinical stages 1, 3 and 5 of chronic kidney disease. Based on the research results, the applicant proposed the technical solution of the present application.

[0051] In order to achieve the above-mentioned purpose, the first aspect of the present invention protects a marker for predicting the clinical stage of chronic kidney disease, wherein the marker is selected from one or more of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, and the clinical stage is stage 1, stage 3 and stage 5. The present invention newly finds that the content of one or more volatile organic compounds in n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane in exhaled breath has significant differences between stage 1, stage 3 and stage 5 of chronic kidney disease progression, and further correlation analysis is performed by 7 volatile organic compounds and renal function indicators, and it is found that the content of the 7 volatile organic compounds is significantly correlated with the estimated glomerular filtration rate, serum creatinine and blood urea nitrogen levels, which is negatively correlated with the estimated glomerular filtration rate and positively correlated with serum creatinine and blood urea nitrogen, so it can be used as a marker for detecting clinical stage 1, stage 3 and stage 5 of chronic kidney disease.

[0052] Preferably, the marker may be a combination of n-octane and 4-heptanone, or a combination of n-octane, 4-heptanone and n-dodecane, or a combination of n-octane, 4-heptanone, n-dodecane and o-xylene, or a combination of n-octane, 4-heptanone, n-dodecane, o-xylene and ethylbenzene, or a combination of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene and m-xylene, or a combination of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane.

[0053] Preferably, the marker is derived from the exhaled breath of a chronic kidney disease patient, preferably, the alveolar exhaled breath.

[0054] In order to achieve the above object, another aspect of the present invention protects the use of a substance for detecting the markers as described above in the preparation of a product for predicting the clinical stage of chronic kidney disease. The clinical stage is stage 1, stage 3 and stage 5.

[0055] Preferably, the substance for detecting a marker includes a substance for detecting the volume or mass concentration of the marker.

[0056] In order to achieve the above object, another aspect of the present invention protects a product for predicting the clinical stage of chronic kidney disease, the product comprising the substance for detecting markers as described above. The clinical stage is stage 1, stage 3 and stage 5.

[0057] Preferably, the product comprises at least one of a kit, a membrane strip, a chip and a detection system. The detection system in the present application is selected from a gas sensor system or a gas chromatograph.

[0058] In order to achieve the above object, another aspect of the present invention protects the use of the markers as described above or the substances for detecting the markers as described above in screening drugs for treating and / or preventing chronic kidney disease.

[0059] In order to achieve the above-mentioned purpose, another aspect of the present invention protects a method for screening the markers mentioned above, wherein the method detects the content of organic matter in samples of patients with clinical stage 1, stage 3 and stage 5 of chronic kidney disease, performs differential analysis on the content of the organic matter, and obtains differential organic matter, which is the marker.

[0060] Preferably, a t-test or z-test is used for difference analysis to obtain a P value, and a P value < 0.05 indicates a marker.

[0061] Preferably, the detection method is gas chromatography-mass spectrometry, and the detection method comprises the following steps:

[0062] 1-1) Enriching exhaled air using adsorbent;

[0063] 1-2) A thermal desorber is used to perform thermal desorption on the adsorbent after enriching the exhaled breath, and then a gas chromatography-mass spectrometer is used for detection. The detection results are qualitatively and quantitatively analyzed, and the types of organic matter are obtained according to the qualitative analysis, and the content of organic matter is obtained according to the quantitative analysis.

[0064] Preferably, the adsorbent is selected from one or more of Tenax, Carbograph, Chromosorb and PoraPak.

[0065] Preferably, in step 1-2), the detection results are qualitatively and quantitatively analyzed according to the NIST14 database and the retention time of the standard substance.

[0066] Preferably, the conditions for thermal desorption are:

[0067] The primary desorption temperature is 280-320°C, and the desorption time is 5-15 min;

[0068] The secondary desorption temperature is 300-340°C and the desorption time is 2-8 min;

[0069] Transmission line temperature: 180~220℃;

[0070] No diversion.

[0071] Preferably, the gas chromatography measurement conditions are:

[0072] Chromatographic column: HP-5ms, length 30m×inner diameter 0.25mm×film thickness 0.25μm;

[0073] Carrier gas: Helium purity ≥ 99.999%;

[0074] Carrier gas pressure: (5~15) psi;

[0075] Heating procedure: initial temperature 30-50°C, maintained for 2-8 min; then raised to 250-290°C at a rate of 5-15°C / min, maintained for 2-8 min.

[0076] Preferably, the mass spectrometry conditions are:

[0077] Transmission line temperature: 220~260℃;

[0078] Ionization method: electron impact (EI);

[0079] Ionization voltage: 60-80 eV;

[0080] Ion source temperature: 210~250℃;

[0081] Quadrupole temperature: 130~170℃;

[0082] Detection mode: qualitative analysis by full scan total ion current (TIC), semi-quantitative analysis by peak area method;

[0083] Mass range: 35~390amu.

[0084] In order to achieve the above object, another aspect of the present invention protects a method for constructing a chronic kidney disease clinical staging prediction model, the construction method comprising:

[0085] 1) constructing a sample data set based on the detection amount of the markers as described above in the subject samples, wherein the subjects include patients with confirmed clinical stage 1, stage 3, and stage 5 chronic kidney disease;

[0086] 2) randomly dividing the sample data set into a test set and a training set, using a machine learning method to learn the training set to obtain the chronic kidney disease clinical staging prediction model.

[0087] In certain embodiments of the present invention, the sample is selected from exhaled air, preferably alveolar exhaled air.

[0088] In certain embodiments of the present invention, the machine learning method includes at least one of a support vector machine algorithm, a nearest neighbor algorithm, and a linear discriminant analysis algorithm.

[0089] In certain embodiments of the present invention, the step 3) is also included in which the chronic kidney disease clinical staging prediction model is tested using a test set.

[0090] In the present invention, the kernel function of the support vector machine algorithm is a polynomial kernel.

[0091] In certain embodiments of the present invention, different expected output values ​​are set according to different categories of samples: the expected output value of samples from patients with stage 1 chronic kidney disease is 1, the expected output value of samples from patients with stage 3 chronic kidney disease is 3, and the expected output value of samples from patients with stage 5 chronic kidney disease is 5.

[0092] In certain embodiments of the present invention, a leave-one-out cross-validation method is used to learn and validate the subject data set. Specifically, for example, 59 subject sample data, each time using the sample data of one subject as the test set, the remaining 58 subjects' exhaled breath marker detection data as the training set, using the training set to train the learning machine model, using the test set data for validation, cycle 59, each test set of subjects is different. After multiple iterations, the model parameters with the highest accuracy are obtained, and the harmonic mean of the accuracy, precision and recall of the model leave-one-out cross-validation under this parameter is the model performance evaluation result.

[0093] In a specific embodiment of the present invention, a machine learning method is used to learn a training set (i.e., the content data of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane in the training set), and then a test set (i.e., the content data of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane in the test set) is used for testing to obtain the chronic kidney disease clinical staging prediction model.

[0094] In the present invention, different machine learning methods were evaluated by accuracy. The results showed that the accuracy of the model constructed by the support vector machine algorithm based on the marker combination was 71.2% to 76.3%; the accuracy of the model constructed by the nearest neighbor element algorithm based on the marker combination was 64.4% to 74.6%; the accuracy of the model constructed by the linear discriminant analysis algorithm based on the marker combination was 66.1% to 69.5%; among them, the model constructed by the support vector machine algorithm based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene had the highest accuracy of 76.3%. In summary, the variation of the model constructed by the present invention is not obvious, the accuracy is relatively high, and the markers of the present invention can accurately predict the clinical stage of chronic kidney disease.

[0095] In the present invention, the model was further evaluated by calculating the harmonic means of precision and recall. The results showed that the harmonic means of precision and recall were both above 50%, indicating that the model of the present invention can specifically identify patients in stage 1, stage 3, and stage 5.

[0096] In order to achieve the above object, another aspect of the present invention protects a system for predicting the clinical stage of chronic kidney disease, the system comprising:

[0097] S1. Obtaining the detection amount data of the markers as described above in the sample of the subject;

[0098] S2. Using the chronic kidney disease clinical staging prediction model obtained by the construction method described above to process the detection data, so as to output the chronic kidney disease clinical staging results.

[0099] In order to achieve the above object, another aspect of the present invention protects a method for predicting the clinical stage of chronic kidney disease, comprising the following steps:

[0100] S1. Obtaining the detection amount data of the markers as described above in the sample of the subject;

[0101] S2. Applying the chronic kidney disease clinical staging prediction model obtained by the construction method described above to process the detection data to output the chronic kidney disease clinical staging result;

[0102] In certain embodiments of the present invention, the sample is exhaled air, preferably alveolar exhaled air.

[0103] In order to achieve the above-mentioned purpose, another aspect of the present invention protects a device, including a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the device performs the prediction method as described above or the construction method as described above.

[0104] In order to achieve the above object, another aspect of the present invention protects a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the prediction method as described above or the construction method as described above.

[0105] The present invention screened 16 organic substances from the exhaled breath organic substances of 21 patients with clinical stage 1 of chronic kidney disease, 17 patients with clinical stage 3 of chronic kidney disease and 21 patients with clinical stage 5 of chronic kidney disease, and found that there were significant differences in n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane between patients with stage 1, stage 3 and stage 5. Further, a detection model for clinical staging of chronic kidney disease was established using these 7 different organic substances. Using this model, the clinical staging of chronic kidney disease in the test set can be distinguished, and the accuracy rate is as high as 76.3%. This shows that the detection model for clinical staging of chronic kidney disease established using 7 volatile organic substances, n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, can effectively distinguish clinical stage 1, stage 3 and stage 5 of chronic kidney disease.

[0106] Figure 1 It is a schematic diagram of the exhaled air collection device of the present invention. The device used in the present invention includes an exhaled air collection device, an adsorption tube, a thermal desorber and a gas chromatograph-mass spectrometer. The specific method of use is to use CO2 to control the exhaled air collection device to collect the alveolar exhaled air of the subject in a Teflon air bag, and then transfer the exhaled air in the air bag to an adsorption tube filled with Tenax TA adsorption material for enrichment, and put the adsorption tube enriched with exhaled air components into the thermal desorber for heating to desorb the exhaled air components, and the desorbed exhaled air components are delivered to a gas chromatograph-mass spectrometer with a chromatographic column of HP-5ms for detection. The following embodiments of the present application use a UNITY-xr TD thermal desorber (Markes International, UK); an 8890 / 5977B gas chromatograph-mass spectrometer (Agilent, USA);

[0107] The specific conditions of thermal desorption are:

[0108] The primary desorption temperature was 300 °C and the desorption time was 10 min;

[0109] The secondary desorption temperature was 320 °C and the desorption time was 3 min;

[0110] Transmission line temperature: 200°C;

[0111] No diversion.

[0112] The specific conditions for gas chromatography determination are:

[0113] Chromatographic column: Chromatographic column: Chromatographic column HP-5ms, length 30m × inner diameter 0.25mm × film thickness 0.25μm;

[0114] Carrier gas: Helium purity ≥ 99.999%;

[0115] Carrier gas pressure: 10psi;

[0116] Heating program: initial temperature 40°C, maintained for 5 min; then increased to 270°C at a rate of 10°C / min, maintained for 5 min.

[0117] The specific conditions for mass spectrometry are:

[0118] Transmission line temperature: 240°C;

[0119] Ionization method: electron impact (EI);

[0120] Ionization voltage: 70 eV;

[0121] Ion source temperature: 230°C;

[0122] Quadrupole temperature: 150℃;

[0123] Detection mode: qualitative analysis by full scan total ion current (TIC), semi-quantitative analysis by peak area method;

[0124] Mass range: 35~390amu.

[0125] All materials, reagents and instruments used in the present invention are well known in the art, but do not limit the implementation of the present invention. Other reagents and equipment well known in the art can be applied to the implementation of the following embodiments of the present invention.

[0126] Example 1 Screening of clinical staging markers for chronic kidney disease

[0127] In this embodiment, the Figure 1 The CO2 controlled exhaled air collection device shown collects the alveolar exhaled air of the subject, and detects the organic matter in the collected alveolar exhaled air by gas chromatography-mass spectrometry, performs qualitative and quantitative analysis on the test results, and screens organic matter with significant differences in content as markers. Including the following:

[0128] Subjects were divided into 3 groups: chronic kidney disease clinical stage 1 group (21 people), chronic kidney disease clinical stage 3 group (17 people) and chronic kidney disease clinical stage 5 group (21 people).

[0129] All subjects rinsed their mouths with pure water before collecting exhaled breath.

[0130] Methods and Design: The exhaled breath of patients with clinical stage 1, 3 and 5 chronic kidney disease was collected. After pretreatment, the exhaled breath was analyzed by gas chromatography-mass spectrometry to obtain the chromatographic peak area of ​​volatile organic compounds. The chromatographic peak area of ​​volatile organic compounds in the exhaled breath of the chronic kidney disease clinical stage 1 group, the chronic kidney disease clinical stage 3 group and the chronic kidney disease clinical stage 5 group was analyzed univariately to screen out the volatile organic compounds with significant differences between the two groups.

[0131] 1.1 Collection of exhaled breath

[0132] Based on the Fowler model, the exhaled gas CO2 concentration is increased by 50% (C50) to distinguish between anatomical dead space exhaled gas and alveolar exhaled gas. When the exhaled gas is lower than C50, the collection device does not collect this part of the exhaled gas; when the exhaled gas is higher than C50, the collection device collects this part of the exhaled gas in a Teflon air bag. The volume of the exhaled gas collection bag is 3L.

[0133] 1.2 Pretreatment and detection of exhaled breath

[0134] The exhaled air of each group was transferred to an adsorption tube filled with Tenax TA adsorbent for enrichment, and then detected by gas chromatography-mass spectrometry through thermal desorption injection to obtain the test results.

[0135] 1.3 Analysis of test results

[0136] According to the NIST14 database and the retention time of the standard substances, the test results obtained in step 1.2 were qualitatively and quantitatively analyzed, and a total of 16 volatile organic compounds in the exhaled breath were identified. The logarithmic values ​​of the chromatographic peak areas of the 16 volatile organic compounds are shown in Table 1.

[0137] Figure 2 : is the chromatogram of exhaled breath in this example.

[0138] The t-test method was used to compare the logarithmic values ​​of the chromatographic peak areas of 16 volatile organic compounds identified between the chronic kidney disease clinical stage 1 group, the chronic kidney disease clinical stage 3 group, and the chronic kidney disease clinical stage 5 group. The volatile organic compounds with significant differences (p<0.05) between the two groups were regarded as markers.

[0139] Table 1

[0140]

[0141]

[0142] A represents p<0.001, B represents p<0.005, and C represents p<0.05vs 1 period, a represents p<0.001, b represents p<0.005, and c represents p<0.05vs 3 periods.

[0143] 1.4 Screening of Volatile Organic Compounds as Biomarkers

[0144] As shown in Table 1, the 16 volatile organic compounds are n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene, n-heptane, n-nonane, n-undecane, n-tridecane, n-tetradecane, n-hexadecane, styrene, heptaldehyde, 2-methyl-2-butanol and butyl acetate.

[0145] As shown in Table 1, the comparison results of the chronic kidney disease clinical stage 1 group and the chronic kidney disease clinical stage 3 group showed that the p values ​​of 8 organic substances, including n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene, n-heptane and 2-methyl-2-butanol, were all less than 0.05. Combined with the average values ​​of their contents, it can be seen that the contents of 8 organic substances in the chronic kidney disease clinical stage 1 group were significantly lower than those in the chronic kidney disease stage 3 group; the p values ​​of 5 organic substances, including 4-heptanone, o-xylene, ethylbenzene, m-xylene and 2-methyl-2-butanol, were in the range of 0.005-0.05; the p values ​​of 2 organic substances, including n-dodecane and n-heptane, were in the range of 0.001-0.005; and the p value of n-octane was less than 0.001. The results showed that among the 16 organic compounds, the content of n-octane had the greatest difference between the two groups, while the contents of n-nonane, n-undecane, n-tridecane, n-tetradecane, n-hexadecane, styrene, heptaldehyde and butyl acetate had no significant differences between the chronic kidney disease clinical stage 1 group and the chronic kidney disease clinical stage 3 group.

[0146] The comparison results between the CKD clinical stage 3 group and the CKD clinical stage 5 group showed that, except for the contents of 2-methyl-2-butanol and butyl acetate, which had no significant differences between the two groups, the contents of the other 14 volatile organic compounds were significantly lower in the CKD clinical stage 3 group than in the CKD clinical stage 5 group, among which the p values ​​of ethylbenzene and n-heptane were 0.005-0.05, the p values ​​of n-octane and n-tridecane were 0.001-0.005, and the p values ​​of 4-heptanone, n-dodecane, o-xylene, m-xylene, n-nonane, n-undecane, n-tetradecane, n-hexadecane, styrene and heptaldehyde were all less than 0.001. This showed that among the 16 organic compounds, the contents of 4-heptanone, n-dodecane, o-xylene, m-xylene, n-nonane, n-undecane, n-tetradecane, n-hexadecane, styrene and heptaldehyde were the most different between the CKD clinical stage 3 group and the CKD clinical stage 5 group.

[0147] The comparison results between the CKD clinical stage 1 group and the CKD clinical stage 5 group showed that, except for the content of butyl acetate, which had no significant difference between the two groups, the contents of the other 15 volatile organic compounds in the CKD clinical stage 1 group were significantly lower than those in the CKD clinical stage 5 group, among which the p value of 2-methyl-2-butanol was in the range of 0.005-0.05, and the p values ​​of the other 14 volatile organic compounds were all less than 0.001. It showed that among the 16 organic compounds, n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene, n-heptane, n-nonane, n-undecane, n-tridecane, n-tetradecane, n-hexadecane, styrene and heptaldehyde had the greatest difference between the CKD clinical stage 1 group and the CKD clinical stage 5 group.

[0148] Based on the comparison results between the above groups, it can be seen that among the 16 volatile organic compounds, only the contents of 7 markers, including n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, gradually increased with the increasing severity of chronic kidney disease, and there were significant differences between the groups. Therefore, these 7 markers were selected for further study.

[0149] Example 2 Correlation analysis between exhaled breath markers and renal function indicators in clinical stages of chronic kidney disease

[0150] In this example, the seven markers obtained in Example 1 were used to perform correlation analysis with renal function indicators such as estimated glomerular filtration rate, serum creatinine and blood urea nitrogen levels, and further, the correlation between the seven markers and chronic kidney disease was verified.

[0151] The serum of the subjects in Example 1 was collected respectively, and the Longhua Hospital affiliated to Shanghai University of Traditional Chinese Medicine was commissioned to measure the creatinine level and blood urea level in the serum, and the estimated glomerular filtration rate was calculated. The contents of the seven markers were correlated with the estimated glomerular filtration rate, serum creatinine and blood urea nitrogen levels, and the correlation coefficient results are shown in Table 2.

[0152] Table 2

[0153]

[0154] A represents p<0.001.

[0155] Correlation coefficient: generally represented by the letter r, used to measure the linear relationship between two variables. The r value is between -1 and 1, r<0 is a negative correlation, and r>0 is a positive correlation. When the correlation between two variables reaches the maximum, r takes a value of -1 or 1. The p value indicates the probability of no correlation, and p<0.05 indicates that the two variables are significantly correlated.

[0156] As can be seen from Table 2, 7 markers including n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane are significantly correlated with estimated glomerular filtration rate, serum creatinine and blood urea nitrogen, among which they are negatively correlated with estimated glomerular filtration rate and positively correlated with serum creatinine and blood urea nitrogen. Therefore, one or a combination of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane screened out by the present invention can be used as a marker for detecting clinical stage 1, stage 3 and stage 5 of chronic kidney disease.

[0157] Example 3 Construction of a chronic kidney disease staging prediction model

[0158] In this example, the contents of two markers, n-octane and 4-heptanone, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0159] 1) The marker contents in the exhaled breath of 59 subjects (21 patients with clinical stage 1 chronic kidney disease, 17 patients with clinical stage 3 chronic kidney disease, and 21 patients with clinical stage 5 chronic kidney disease) were detected and obtained respectively, and the detection data of the marker contents in the exhaled breath of 1 subject were used as the test set each time, and the detection data of the marker contents in the exhaled breath of the remaining 58 subjects were used as the training set, and the data of the training set were used to train the learning machine model, and the data of the test set were used for verification, and the cycle was 59, and the subjects of each test set were different; the input features of the training set and the test set were the contents of n-octane and 4-heptanone.

[0160] The data in the training set were learned using the vector machine, nearest neighbor algorithm, and linear discriminant analysis algorithm. Different expected output values ​​were set according to different categories of samples: the expected output value of patients with stage 1 chronic kidney disease was 1, the expected output value of patients with stage 3 chronic kidney disease was 3, and the expected output value of patients with stage 5 chronic kidney disease was 5. After multiple iterations, the model parameters with the highest accuracy were obtained. The accuracy of the model under the parameters and the harmonic mean of the precision and recall of the leave-one-out cross-validation were the model performance evaluation results.

[0161] The model was evaluated by calculating the accuracy and the harmonic mean of precision and recall. The results are shown in Tables 3 and 4.

[0162] Accuracy: The ratio of correctly classified samples to the total number of samples.

[0163] Precision: The ratio of correctly classified positive data to the predicted positive data.

[0164] Recall rate: The ratio of correctly classified positive data to actual positive data.

[0165] The harmonic mean of precision and recall = 2 / (1 / precision + 1 / recall).

[0166] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 66.1% to 71.2%, among which the chronic kidney disease staging prediction model constructed using the support vector machine algorithm has the highest accuracy of 71.2%.

[0167] Example 4 Construction of a chronic kidney disease staging prediction model

[0168] In this example, the contents of three markers, n-octane, 4-heptanone and n-dodecane, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0169] The difference from Example 3 is that the input features of the training set and the test set are the contents of n-octane, 4-heptanone and n-dodecane, and the rest are the same as Example 3. The results are shown in Tables 3 and 4.

[0170] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 67.8% to 74.6%, among which the chronic kidney disease staging prediction model constructed by the support vector machine algorithm and the nearest neighbor element algorithm has the highest accuracy, both of which are 74.6%.

[0171] Example 5 Construction of a chronic kidney disease staging prediction model

[0172] In this example, the contents of four markers, namely n-octane, 4-heptanone, n-dodecane and o-xylene, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0173] The difference from Example 3 is that the input features of the training set and the test set are the contents of n-octane, 4-heptanone, n-dodecane and o-xylene, and the rest are the same as Example 3. The results are shown in Tables 3 and 4.

[0174] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 67.8% to 76.3%, among which the chronic kidney disease staging prediction model constructed by the support vector machine algorithm has the highest accuracy of 76.3%.

[0175] Example 6 Construction of a chronic kidney disease staging prediction model

[0176] In this example, the contents of five markers, namely n-octane, 4-heptanone, n-dodecane, o-xylene and ethylbenzene, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0177] The difference from Example 3 is that the input features of the training set and the test set are the contents of n-octane, 4-heptanone, n-dodecane, o-xylene and ethylbenzene, and the rest are the same as Example 3. The results are shown in Tables 3 and 4.

[0178] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 66.1% to 72.9%, among which the chronic kidney disease staging prediction model constructed using the support vector machine algorithm has the highest accuracy of 72.9%.

[0179] Example 7 Construction of a chronic kidney disease staging prediction model

[0180] In this example, the contents of six markers, namely n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene and m-xylene, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0181] Same as Example 3, except that the input features of the training set and the test set are the contents of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene and m-xylene, and the rest are the same as Example 3. The results are shown in Tables 3 and 4.

[0182] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 66.1% to 72.9%, among which the chronic kidney disease staging prediction model constructed by the support vector machine algorithm and the linear discriminant analysis algorithm has the highest accuracy of 72.9%.

[0183] Example 8 Construction of a chronic kidney disease staging prediction model

[0184] In this example, the contents of seven markers, namely n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, were used as input features, and different machine learning models were used to construct a chronic kidney disease staging prediction model, including the following:

[0185] The difference from Example 3 is that the input features of the training set and the test set are the contents of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, and the rest are the same as Example 3. The results are shown in Tables 3 and 4.

[0186] As can be seen from Table 3, in this embodiment, the accuracy of the chronic kidney disease staging prediction model constructed using the three models is 64.4% to 72.9%, among which the chronic kidney disease staging prediction model constructed using the support vector machine algorithm has the highest accuracy of 72.9%.

[0187] Table 3

[0188]

[0189] Table 4

[0190]

[0191] Among them, marker combination 1-2 in Table 3 and Table 4 is the composition numbered 1 and 2 in Table 1, that is, the combination of n-octane and 4-heptanone; 1-3 is the composition numbered 1, 2 and 3 in Table 1, that is, the combination of n-octane, 4-heptanone and n-dodecane; the same applies to 1-4, 1-5, 1-6 and 1-7.

[0192] As shown in Table 3, the accuracy of the model constructed by the support vector machine algorithm based on the marker combination is 71.2% to 76.3%, among which the model constructed based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene has the highest accuracy; the accuracy of the model constructed by the nearest neighbor element algorithm based on the marker combination is 64.4% to 74.6%, among which the models constructed based on the combination of n-octane + 4-heptanone + n-dodecane and n-octane + 4-heptanone + n-dodecane + o-xylene have the highest accuracy; the accuracy of the model constructed by the linear discriminant analysis algorithm based on the marker combination is 66.1% to 69.5%, among which the models constructed based on the combination of n-octane + 4-heptanone and n-octane + 4-heptanone + n-dodecane + o-xylene + ethylbenzene + m-xylene have the highest accuracy, both of which are 69.5%. Among all the models, the model constructed by the support vector machine algorithm based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene has the highest accuracy, which is 76.3%.

[0193] As can be seen from Table 4, the harmonic mean of the 1-phase precision and recall of the model constructed by the support vector machine algorithm based on the marker combination is 70.6%~78.9%, among which the model constructed based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene has the highest accuracy; the harmonic mean of the 3-phase precision and recall of the model constructed by the support vector machine algorithm based on the marker combination is 65.3%~73.9%, among which the model constructed based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene has the highest accuracy; the harmonic mean of the 5-phase precision and recall of the model constructed by the support vector machine algorithm based on the marker combination is 76.5%~80.0%, among which the model constructed based on the combination of n-octane + 4-heptanone has the highest accuracy. The harmonic mean of the 1-phase precision and recall of the model constructed by the nearest neighbor algorithm based on the marker combination was 63.2% to 78.9%, among which the model constructed based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene had the highest accuracy; the harmonic mean of the 3-phase precision and recall of the model constructed by the nearest neighbor algorithm based on the marker combination was 51.3% to 71.1%, among which the model based on the combination of n-octane + 4-heptanone + n-dodecane had the highest accuracy; the harmonic mean of the 5-phase precision and recall of the model constructed by the nearest neighbor algorithm based on the marker combination was 70.3% to 80.0%, among which the model constructed based on the combination of n-octane + 4-heptanone + n-dodecane had the highest accuracy. The harmonic mean of the 1-phase precision and recall of the models constructed based on the marker combination using the linear discriminant analysis algorithm was 66.7%-69.8%, among which the models based on the combination of n-octane + 4-heptanone, n-octane + 4-heptanone + n-dodecane, n-octane + 4-heptanone + n-dodecane + o-xylene, n-octane + 4-heptanone + n-dodecane + o-xylene + ethylbenzene + m-xylene and n-octane + 4-heptanone + n-dodecane + o-xylene + ethylbenzene + m-xylene + n-heptane had the highest accuracy, all of which were The harmonic mean of the 3-phase precision and recall of the model constructed by the linear discriminant analysis algorithm based on the marker combination was 55.6%-63.2%, among which the model based on the combination of n-octane + 4-heptanone had the highest accuracy; the harmonic mean of the 5-phase precision and recall of the model constructed by the linear discriminant analysis algorithm based on the marker combination was 73.7%-76.9%, among which the model based on the combination of n-octane + 4-heptanone + n-dodecane + o-xylene + ethylbenzene + m-xylene had the highest accuracy.

[0194] It can be seen that the present invention has found significant differences in n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane in the exhaled breath of patients in stages 1, 3 and 5 through detection. Furthermore, it was found that seven organic substances, namely n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane, were significantly correlated with the estimated glomerular filtration rate, serum creatinine and blood urea nitrogen. Therefore, they can be used as markers to distinguish between stages 1, 3 and 5 chronic kidney disease.

[0195] The present invention also provides a chronic kidney disease clinical staging prediction model, which uses these 7 markers to establish a chronic kidney disease clinical staging detection model, and evaluates the model using accuracy and the harmonic mean of precision and recall. The accuracy is as high as 76.3%, and the harmonic means of precision and recall are both higher than 50%, indicating that the model of the present invention can specifically identify stage 1, stage 3 and stage 5 patients.

[0196] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A marker for predicting the clinical stage of chronic kidney disease, It is characterized in that The marker is derived from the exhaled breath of patients with chronic kidney disease; the marker is selected from one or more of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane; and includes at least one of the following features: 1) Combination of n-octane and 4-heptanone; 2) a combination of n-octane, 4-heptanone and n-dodecane; 3) a combination of n-octane, 4-heptanone, n-dodecane and o-xylene; 4) a combination of n-octane, 4-heptanone, n-dodecane, o-xylene and ethylbenzene; 5) a combination of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene and m-xylene; 6) A combination of n-octane, 4-heptanone, n-dodecane, o-xylene, ethylbenzene, m-xylene and n-heptane.

2. Use of a substance for detecting the marker as claimed in claim 1 in the preparation of a product for predicting the clinical stage of chronic kidney disease.

3. The use according to claim 2, It is characterized in that The substance for detecting a marker includes a substance for detecting the volume or mass concentration of the marker.

4. A product for predicting the clinical stage of chronic kidney disease, It is characterized in that The product comprises the substance for detecting a marker for use as claimed in claim 2 or 3.

5. The product according to claim 4, It is characterized in that The product includes at least one of a test kit, a membrane strip, a chip and a detection system.

6. Use of the marker according to claim 1 or the product according to claim 4 or 5 in screening drugs for treating and / or preventing chronic kidney disease.

7. A method for constructing a prediction model for clinical staging of chronic kidney disease. It is characterized in that The construction method comprises: 1) constructing a sample data set based on the detection amount of the marker as claimed in claim 1 in the subject sample, wherein the subject includes patients with confirmed clinical stage 1, stage 3 and stage 5 chronic kidney disease; 2) Randomly dividing the sample data set into a test set and a training set, using a machine learning method to learn the training set to obtain the chronic kidney disease clinical staging prediction model.

8. The construction method according to claim 7, It is characterized in that The sample is selected from exhaled breath; And / or, the machine learning method is selected from at least one of a support vector machine algorithm, a nearest neighbor algorithm, and a linear discriminant analysis algorithm; And / or, the method further includes step 3) testing the chronic kidney disease clinical staging prediction model using a test set.

9. A method for predicting the clinical stage of chronic kidney disease. It is characterized in that The steps include: S1. Obtaining the detection amount data of the marker according to claim 1 in the sample of the subject; S2. Using the chronic kidney disease clinical staging prediction model obtained by the construction method as described in claim 7 or 8 to process the detection data to output the chronic kidney disease clinical staging results.

10. A system for predicting the clinical stage of chronic kidney disease. It is characterized in that The system comprises: A data module, used to obtain the detection amount data of the marker according to claim 1 in the sample of the subject; An evaluation module processes the detection data using the chronic kidney disease clinical staging prediction model obtained by the construction method according to claim 7 or 8 to output the chronic kidney disease clinical staging results.

11. A device comprising a processor and a memory, the memory being configured to store a computer program, It is characterized in that The processor is configured to execute the computer program stored in the memory, so that the device executes the prediction method according to claim 9 or the construction method according to claim 7 or 8.

12. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the prediction method according to claim 9 or the construction method according to claim 7 or 8 is implemented.

Images