Metabolic markers for children autism and screening method and application thereof

Abstract

The invention provides a group of metabolic markers for children autism as well as a screening method and application thereof, and belongs to the technical field of biology, the metabolic markers comprise methionine, citrulline, phenylalanine, acetyl carnitine, decadiene carnitine and myristoyl carnitine, the screening method comprises the following steps: (1) collecting dried blood spot samples of autism children and normal control children; (2) detecting metabolites of the dried blood spot sample to obtain data; and (3) analyzing and screening the data to obtain the metabolic marker. The metabolic marker obtained by the technical scheme provided by the invention is used for auxiliary diagnosis of children autism, and the purposes of convenience in sampling and good detection and judgment effects are achieved.

Description

Technical Field

[0001] The present invention relates to the field of biotechnology, and in particular to a group of metabolic markers for childhood autism, a screening method and an application thereof. Background Art

[0002] Autism Spectrum Disorder (ASD) is a group of neuropsychological developmental disorders that begin in early childhood and are characterized by social interaction and communication disorders, repetitive stereotyped behaviors, and narrow interests. In the past 20 years, the prevalence of autism in children has increased year by year, and the number of ASD patients worldwide is at least 78 million. Moreover, 54% to 70% of autistic patients often have intellectual disabilities or other mental disorders, and the prognosis is poor, which brings heavy economic and social burdens to the patients' families and society.

[0003] At present, the main basis for the diagnosis of ASD is the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) and the Autism Diagnostic Observation Scale (ADOS), as well as necessary intellectual or developmental diagnostic tests, psychological assessments, physical and nervous system tests, and some necessary auxiliary tests. However, this diagnostic method relies on experienced clinicians and has certain defects: first, it is highly subjective and requires clinicians to accurately grasp the objective medical history and fully understand the mental and psychological development of the child, otherwise it is easy to be confused with other mental illnesses; second, the test is time-consuming, and it takes a year from the initial diagnosis to the final diagnosis; finally, questionnaire surveys are difficult for children under 3 years old. Therefore, there is an urgent need to find a cost-effective, objective and practical biomarker strategy that can assist in the diagnosis of ASD based on the above medical history.

[0004] Metabolomics is a powerful tool for studying metabolic processes in organisms because it can keenly capture the interactions between genes, diet and environmental factors, thereby shaping the unique metabolic characteristics of an individual. Based on this concept, metabolomics analysis can reveal key biological indicators and help identify disturbances in biological processes that lead to ASD. For example, the patent with publication number CN117280215A and the name Diagnosis of Autism Spectrum Disorders discloses that some metabolic markers such as ribitol, lyxonate, erythritol or ribose and uric acid have specific differences compared with typical development controls, that is, the subjects are diagnosed with ASD. However, the sampling and detection stability of metabolic markers in the above scheme still needs to be improved.

[0005] Therefore, it is necessary to provide a group of metabolic markers for childhood autism with good stability, as well as screening methods and applications, in order to solve the problems existing in the above-mentioned prior art. Summary of the invention

[0006] In view of this, the present invention provides a group of metabolic markers for childhood autism, a screening method and an application thereof. The obtained metabolic markers are applied to the auxiliary diagnosis of childhood autism, achieving the purpose of easy sampling and good detection and discrimination effect.

[0007] To achieve the above object, the present invention provides a group of metabolic markers for childhood autism, wherein the metabolic markers include methionine, citrulline, phenylalanine, acetylcarnitine, decadienylcarnitine and tetradecanoylcarnitine.

[0008] To achieve the above object, the present invention also provides a method for screening the above-mentioned metabolic markers of childhood autism, comprising the following steps: (1) Collecting dried blood samples from autistic patients and normal control children; (2) Detect metabolites in dried blood samples and obtain data; (3) Analyze and screen the data to obtain metabolic markers.

[0009] Optionally, the analysis and screening of the data in (3) includes the following steps: S1. The samples of the autism patients group and the normal control children group were divided into a discovery set and an external validation set. The Wilcoxon rank sum test and logistic regression in non-parametric tests were used to screen out the metabolites with significant differences in the discovery set and the external validation set, and the intersection was taken to retain the metabolites that appeared together and had the same change direction; S2. Based on the discovery set data, biomarkers were further screened. To evaluate the model performance of the algorithm, the autism group and the control group data of the discovery set were randomly divided into a training set and a validation set through machine learning. The optimal combination of metabolites was screened based on the Lasso regression model and feature reduction. The AUC value of the screened metabolite combination model was calculated based on the multivariate and multi-classification model. S3. Based on the optimal variable combination and statistical model, the diagnostic efficacy of the metabolite combination was further verified by calculating the AUC value of the metabolite combination in external data.

[0010] Optionally, the data ratio of the training set and the validation set in S2 is 8:2.

[0011] Optionally, when calculating the AUC value in S2, the segmentation process is repeated 5 times, the average value of the AUC value is obtained, and its 95% confidence interval is obtained.

[0012] Optionally, the dried blood sample in (1) is collected in the early morning on an empty stomach and stored in a -80°C refrigerator.

[0013] Optionally, liquid chromatography-mass spectrometry is used for detection in (2).

[0014] Optionally, the operating parameters of the liquid chromatography-mass spectrometry instrument include a mobile phase of 80% methanol-water solution, a flow rate of the mobile phase of 0.2 mL / min, which is reduced to 0.1 mL / min within 0.08 min, maintained for 1.5 min, then increased to 0.2 mL / min within 0.1 min, and maintained for 0.5 min; the mass spectrometry ion spray voltage is set to 4.5 kV. The auxiliary gas temperature is maintained at 350°C, and the ion source is an electrospray ion source in positive ion mode.

[0015] To achieve the above object, the present invention also provides a use of the above-mentioned metabolic marker for childhood autism, wherein the metabolic marker is contained in a blood sample of a patient.

[0016] Optionally, the blood sample includes a dried blood sample, a serum sample and a plasma sample.

[0017] The above technical solution of the present invention includes at least the following beneficial effects: The present invention provides a method for screening metabolic markers of childhood autism. By combining children's dry blood samples with metabolomics research, it is found that compared with normal children, there are obvious differences in the metabolic markers of methionine (MET), citrulline (CIT), phenylalanine (PHE), acetylcarnitine (C2), decadienylcarnitine (C10:2) and tetradecanoylcarnitine (C14) in the peripheral blood of children with autism. Then, through machine learning screening, it is determined that MET, CIT, PHE, C2, C10:2 and C14 in the dry blood samples can be used as a potential marker combination for discriminating and diagnosing children with autism and healthy children, and the discriminating and diagnosing effect has a high accuracy rate, which can provide basic data support for early screening, auxiliary diagnosis and research intervention measures of autism.

[0018] The dried blood film used in the present invention is a minimally invasive blood collection method and can be used as an alternative to whole blood samples collected by venipuncture or arterial sampling. It is convenient for sampling, has high stability, requires less blood, is easy to store and transport, and can reduce the risk of infection. It can perform high-throughput, reliable and stable determination of multiple analytes. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of the process data of obtaining six metabolic markers by screening in Example 2 of the present invention; Figure 2 This is a ROC analysis curve diagram of the training set in Example 3 of the present invention; Figure 3 This is a ROC analysis curve diagram of the test set in Example 3 of the present invention; Figure 4 This is a ROC analysis curve diagram of the test set in Example 4 of the present invention. DETAILED DESCRIPTION

[0020] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the following will be combined with the appended drawings of the embodiments of the present invention. Figures 1 to 4 , the technical scheme of the embodiment of the present invention is clearly and completely described. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0021] Nouns used in the present invention: The Receiver Operating Characteristic (ROC) curve is a curve drawn with the false positive rate (1-specificity, i.e., misdiagnosis rate) as the horizontal axis and the sensitivity (true positive rate) as the vertical axis. The area under the curve (AUC) can determine the effectiveness of the diagnostic test. When AUC>0.5, the closer the AUC is to 1, the better the diagnostic effect.

[0022] Confidence Interval (CI): A confidence interval is an estimated interval of a population parameter constructed from a sample statistic. In statistics, a confidence interval of a probability sample is an interval estimate of a population parameter of the sample. The confidence interval shows the degree to which the true value of the parameter has a certain probability of falling around the measurement result, and it gives the degree of credibility of the measured value of the measured parameter.

[0023] Example 1 Subject Condition and Sample Collection Written informed consent was obtained from all subjects before participating in the study.

[0024] Inclusion and exclusion criteria were as follows: Diagnostic criteria for autism: (1) Children aged 1 to 6 years old; (2) Diagnosed as autism by two experienced physicians from a child psychology and behavioral center based on the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders.

[0025] Exclusion criteria for autism: Children with other developmental disorders or psychiatric illnesses (e.g., Rett syndrome, cerebral palsy, or chronic epileptic seizures).

[0026] The normal control group children were 1-6 years old and had no neurological diseases, systemic medical diseases, signs of any developmental disorders or any family history of psychiatric disorders.

[0027] The study passed the ethical review of the Ethics Committee. In accordance with the principle of informed consent of the subjects, the guardians of all participating research subjects signed the relevant informed consent forms and filled out the relevant questionnaires before sample collection.

[0028] A total of 395 autistic patients and 238 normal control children were collected from two medical centers. All the dried blood samples were collected in the early morning when fasting, and all the collected samples were stored in a -80℃ refrigerator.

[0029] Example 2 Detection of small molecule metabolites in dried blood samples and screening of metabolic markers Screening of metabolic markers: (1) Instruments and reagents: Two sets of liquid phase tandem mass spectrometers were used, including the American API3200MD tandem mass spectrometer, the Japanese Shimadzu SIL-20AC high performance liquid chromatograph autosampler, the American Waters Acquity TMTQD tandem mass spectrometer, the Waters2777C autosampler, and the Waters1525 high performance liquid chromatograph; the reagents used were the non-derivatized multiple amino acid and carnitine determination kits from the Finnish PerkinElmer company.

[0030] (2) Sample preparation and quality control: Use a semi-automatic puncher (specification 1926-071) to punch 3.2 mm blood smears in the wells of a V-shaped cut-bottom microplate. Add 100 μL of the working solution containing the isotope-labeled internal standard to each well at a volume ratio of 1:110. Seal the plate with an adhesive plastic film and gently shake the plate at room temperature for 45 minutes. After the incubation, tear off the adhesive tape from the plate, then transfer 75 μL of the mixture to a new 96-well V-bottom plate, and finally seal the plate with an aluminum foil cover and test it on the machine. Each V-shaped plate specimen is preceded and followed by a high- and low-level quality control product that comes with the kit, which is tested under the same conditions as the specimen; the quality control products are qualified according to the dry blood smear quality evaluation standards of the Clinical Laboratory Center of the Ministry of Health.

[0031] (3) Instrument parameters: A mixture of 80% methanol and water was used as the mobile phase. The elution program started at 0.2 mL / min, then decreased to 0.1 mL / min within 0.08 min and remained constant for 1.5 min. Next, the flow rate was increased to 0.2 mL / min within 0.1 min and maintained for 0.5 min. The mass spectrometer ion spray voltage was set to 4.5 kV. The auxiliary gas temperature was maintained at 350 °C, and the ion source was an electrospray ion source in positive ion mode. The raw data were collected by Analyst v1.6.0 software. ChemoView 2.0.2 software was used for quantification of individual metabolites.

[0032] (4) Data analysis: In the first step, the samples of the autism group and the normal control group were divided into a discovery set and an external validation set according to different sampling centers. The Wilcoxon rank sum test and logistics regression in the non-parametric test were used to screen the metabolites with significant differences in the discovery set and the external validation set, and the intersection was taken to retain the metabolites with the same change direction. In the second step, based on the discovery set data, biomarkers were further screened. In order to evaluate the model performance of the algorithm, the data of the autism group and the control group in the discovery set were randomly divided into a training set and a validation set (in proportion of 80% and 20%) by machine learning. The best combination of metabolites was screened based on the Lasso regression model and feature reduction (REF and XGBOOST). Based on the multivariate multi-classification model, the AUC value corresponding to the ROC of the screened metabolite combination model was calculated. In order to avoid sampling bias, the segmentation process was repeated 5 times, the average value of the AUC value was obtained, and its 95% confidence interval was obtained. In the third step, based on the best variable combination and statistical model, the AUC value of the metabolite combination was further calculated in the external data to verify the diagnostic efficacy of the metabolite combination.

[0033] Results Analysis 1) Subjects: All samples in the discovery set and external validation set are actual samples collected from two medical centers, and the discovery set and external validation set are samples from different centers. The specific number of people in the discovery set: 314 dry blood samples in the autism group and 157 dry blood samples in the normal control group; the number of people in the external validation group: 81 dry blood samples in the autism group and 81 dry blood samples in the normal control group; the discovery set samples were first used to complete the screening of metabolic markers and build the diagnostic model, and finally the validation group samples were used in the validation step.

[0034] Table 1 Subjects

[0035] 2) The metabolomic data of the control group and autism group samples in the above discovery set and external validation set were compared between groups and subjected to logistic regression analysis. The significantly different metabolites were screened out under the condition of P < 0.001, and then the intersection was taken, and finally a total of 21 differential metabolites were obtained.

[0036] 3) For the above 21 differential metabolites, the ASD and control group samples in the discovery set were randomly divided into a training set and a validation set (80% and 20% respectively) by machine learning, and a total of 6 significantly different metabolites were screened based on the Lasso regression model and feature reduction analysis ( Figure 1 and Table 2).

[0037] Table 2 Six metabolic markers with significant differences

[0038] Example 3 Construction of a diagnostic model for distinguishing autistic children from healthy children In order to evaluate the effect of the screened metabolic markers in distinguishing health from autism, multivariate multi-classification ROC analysis was performed again for the six significantly different metabolic markers. The discovery set sample data was randomly divided into a training set (training) and a test set (test) in a ratio of 80% and 20%. Among them, the training set is used to build and train the machine learning classification model, and the test set is used to evaluate the discriminative ability of the trained model. XGBOOST is used to randomly iterate 1000 times, and the evaluation model is constructed by statistically calculating the average accuracy of the final model. The results show that the training set AUC = 1.0; the test set AUC = 0.991 (see Table 3, Figure 2 and Figure 3 ), indicating that the diagnostic model is efficient.

[0039] The ROC curve is a method to study the relationship between model sensitivity and specificity, with sensitivity as the ordinate and 1-specificity as the abscissa. The evaluation basis is to compare the size of the area under the curve (AUC). When AUC is greater than 0.5, the closer AUC is to 1, the better the model performance is, indicating that the diagnostic effect is better. If it is less than 0.5, it means that the accuracy of the model is poor. The specific operation is: after the discovery set data is randomly divided into a training set (training) and a test set (test), 80% of the training set is combined with machine learning to build a model, and the remaining 20% ​​of the test set is used as an unknown sample and put into the diagnostic model constructed by the above training set for blind testing to obtain the ROC analysis curve.

[0040] Table 3 Comparison of the effect of 6 metabolite combinations in assisting diagnosis of autistic children and normal healthy children

[0041] Example 4 Model validation using external validation set dried blood samples The research objects of this embodiment are the same as those of embodiment 1, and the detection and analysis methods of embodiment 2 are the same. In order to verify the effectiveness of the constructed diagnostic model, the samples of the validation group are taken as unknown samples and put into the above-constructed diagnostic model for verification.

[0042] Using the above 6 significantly different metabolic markers, the same XG-BOOST model was used to perform multivariate multi-classification ROC analysis. The results showed that AUC = 0.970 (see Figure 4 ). The above results show that the established prediction model also has a good diagnostic effect in the external validation group.

[0043] The above are preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A set of metabolic markers for childhood autism, characterized in that: The metabolic markers include methionine, citrulline, phenylalanine, acetylcarnitine, decadienylcarnitine and tetradecanoylcarnitine.

2. A method for screening metabolic markers for childhood autism as claimed in claim 1, characterized in that: The following steps are involved: (1) Collecting dried blood samples from children with autism and normal control children; (2) Detect metabolites in dried blood samples and obtain data; (3) Analyze and screen the data to obtain metabolic markers.

3. The method for screening metabolic markers for childhood autism according to claim 2, characterized in that: The analysis and screening of the data in (3) includes the following steps: S1. The samples of the autistic children group and the normal control children group were divided into a discovery set and an external validation set. The Wilcoxon rank sum test and logistic regression in non-parametric tests were used to screen out the metabolites with significant differences in the discovery set and the external validation set, respectively, and the intersection was taken to retain the metabolites that appeared together and had the same change direction; S2. Based on the discovery set data, biomarkers were further screened. To evaluate the model performance of the algorithm, the autism group and the control group data of the discovery set were randomly divided into a training set and a validation set through machine learning. The optimal combination of metabolites was screened based on the Lasso regression model and feature reduction. The AUC value of the screened metabolite combination model was calculated based on the multivariate and multi-classification model. S3. Based on the optimal variable combination and statistical model, the diagnostic efficacy of the metabolite combination was further verified by calculating the AUC value of the metabolite combination in external data.

4. The method for screening metabolic markers for childhood autism according to claim 3, characterized in that: The data ratio of the training set and the validation set in S2 is 8:

2.

5. The method for screening metabolic markers for childhood autism according to claim 3, characterized in that: When calculating the AUC value in S2, the segmentation process was repeated 5 times, the average value of the AUC value was obtained, and a 95% confidence interval was obtained.

6. The method for screening metabolic markers for childhood autism according to claim 2, characterized in that: The dried blood sample sampling condition in (1) is to collect the sample in the early morning on an empty stomach and store it in a -80°C refrigerator.

7. The method for screening metabolic markers for childhood autism according to claim 2, characterized in that: The detection in (2) is performed using a liquid chromatography-mass spectrometer.

8. The method for screening metabolic markers for childhood autism according to claim 7, characterized in that: The working parameters of the liquid chromatography-mass spectrometry instrument include that the mobile phase is 80% methanol-water solution, the flow rate of the mobile phase is 0.2 mL / min, which is reduced to 0.1 mL / min within 0.08 min, maintained for 1.5 min, then increased to 0.2 mL / min within 0.1 min, and maintained for 0.5 min; the mass spectrometry ion spray voltage is set to 4.5 kV. The auxiliary gas temperature is maintained at 350°C, and the equipped ion source is an electrospray ion source in positive ion mode.

9. A use of the metabolic marker for childhood autism as claimed in claim 1, characterized in that: The metabolic markers are contained in a blood sample of the patient.

10. The use of metabolic markers for childhood autism according to claim 9, characterized in that: The blood samples include dried blood samples, serum samples and plasma samples.

Images