Amino acid biomarkers and their applications in wild and farmed silver carp

Abstract

The present application relates to the technical field of silver carp identification or confirmation, and specifically to amino acid biomarkers and applications of wild and farmed silver carp. The biomarkers include at least one of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid. Biomarkers can be accurately used to distinguish wild and farmed silver carp, provide scientific basis and technical support for relevant law enforcement departments, strengthen the supervision and crackdown on illegal fishing in the Yangtze River Basin, reduce overfishing of wild resources, and provide a technical basis for better management and protection of silver carp resources.

Description

Technical Field

[0001] The present application relates to the technical field of silver carp identification or confirmation, and specifically to amino acid biomarkers of wild and farmed silver carp and their applications. Background Art

[0002] Silver carp (Hypophthalmichthys molitrix) is an important aquaculture species in my country, mainly distributed in the Yangtze River Basin. Due to water pollution, overfishing and other factors, the fish biodiversity in the Yangtze River Basin is constantly decreasing. Therefore, my country has fully implemented the "Ten-Year Fishing Ban in the Yangtze River" decision since 2020, cracking down on illegal fishing in the Yangtze River Basin and illegal sales of Yangtze River catches in the market. However, it is currently difficult to distinguish between wild and farmed aquatic products in the Yangtze River Basin in China, and thus it is difficult to meet the requirements of the "Yangtze River Fishing Ban" policy. Therefore, the development of identification and confirmation technology for wild and farmed aquatic products, the exploration of markers for distinguishing wild and farmed silver carp, and the promotion and application of these markers are reliable means to support the "Yangtze River Fishing Ban" policy. Summary of the invention

[0003] In view of the problem that the biological traits of wild and farmed silver carp are similar but the identification technology is lacking, this application uses ultra-high throughput liquid chromatography-mass spectrometry for the first time to collect muscle samples from different wild and farmed silver carp groups, conducts metabolomics analysis, explores the common differences and changes in metabolites of wild and farmed silver carp, and screens out amino acid biomarkers: proline (CAS: 609-36-9), valine (CAS: 7004-03-7), leucine (CAS: 3588- 60-1), Nε-formyl-L-lysine (CAS: 1190-48-3), γ-aminobutyric acid (CAS: 56-12-2), N-formylglycine (CAS: 2491-15-8), N-acetyl-L-lysine (CAS: 1946-82-3), creatinine (CAS: 60-27-5), DL-2-aminooctanoic acid (CAS: 644-90-6), 5-aminolevulinic acid (CAS: 106-60-5). The results of liquid chromatography-mass spectrometry were used for statistical and logistic regression analysis, and a logistic regression analysis model was obtained. The source of the test samples was determined according to the logistic regression model, which can accurately distinguish wild and farmed samples with good sensitivity and specificity. In addition, the biomarkers and uses provided in this application can provide scientific basis and technical support for relevant law enforcement departments, strengthen the supervision and crackdown on illegal fishing in the Yangtze River Basin, reduce overfishing of wild resources, and provide a technical basis for better management and protection of silver carp resources.

[0004] To this end, the embodiments of the present application disclose at least the following technical solutions:

[0005] In a first aspect, the embodiments disclose a biomarker comprising at least one of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid.

[0006] In an embodiment of the first aspect, the biomarker comprises at least one of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid.

[0007] In an embodiment of the first aspect, the biomarker comprises at least one of proline, valine, and leucine.

[0008] In an embodiment of the first aspect, the biomarker comprises at least one of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0009] In an embodiment of the first aspect, the biomarker comprises at least one of DL-2-aminooctanoic acid and 5-aminolevulinic acid.

[0010] In the second aspect, the embodiments disclose biomarkers for distinguishing wild silver carp from farmed silver carp, which contain at least one of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid.

[0011] In an embodiment of the second aspect, the biomarkers comprise proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid.

[0012] In a third aspect, the embodiments disclose a biomarker for distinguishing wild silver carp from farmed silver carp, which contains at least one of proline, valine, and leucine.

[0013] In an embodiment of the third aspect, the biomarkers comprise proline, valine, leucine.

[0014] In a fourth aspect, the embodiments disclose a biomarker for distinguishing wild silver carp from farmed silver carp, which contains at least one of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0015] In an embodiment of the fourth aspect, the biomarkers comprise Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0016] In a fifth aspect, the embodiments disclose a biomarker for distinguishing wild silver carp from farmed silver carp, which contains at least one of DL-2-aminocaprylic acid and 5-aminolevulinic acid.

[0017] In an embodiment of the fifth aspect, the biomarkers comprise DL-2-aminooctanoic acid, 5-aminolevulinic acid.

[0018] In a sixth aspect, the embodiments disclose a reagent composition comprising a reagent for detecting the biomarker of any one of the first to fifth aspects.

[0019] In an embodiment of the sixth aspect, the reagent comprises a reagent for extracting metabolites from a test sample. In some embodiments, the reagent comprises steel beads and a methanol-water solution (4:1, v / v).

[0020] In the seventh aspect, the embodiments disclose the biomarkers described in any one of the first to fifth aspects and / or the use of the biomarkers described in the first to fifth aspects for detecting the biomarkers in distinguishing wild and farmed silver carp.

[0021] In an eighth aspect, an embodiment discloses a method for distinguishing wild silver carp from farmed silver carp.

[0022] In an embodiment of the eighth aspect, the method includes a method for testing the content of each biomarker in the biomarker described in the second aspect in a test sample. In some embodiments, the test sample is an upper back muscle sample of silver carp. In some embodiments, each biomarker includes proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid.

[0023] In an embodiment of the eighth aspect, the method further includes: obtaining a first logistic regression model and a first threshold; obtaining multiple standardized values ​​of the contents of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the first logistic regression model to obtain a first wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp based on the magnitude of the first wild probability and the first threshold. In some embodiments, the first logistic regression model is: X=-23.428×A+38.212×B+28.738×C-63.957×D-124.040×E-57.712×F+50.106×G+47.266×H+15.842×I+9.120×J+665.268; first wild probability=1 / (1+e－X ); wherein A, B, C, D, E, F, G, H, I, and J are respectively the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid, and the first wild probability represents the probability that the test sample is wild silver carp. In some embodiments, if the first wild probability is greater than the first threshold value, the test sample is wild silver carp. In some embodiments, if the first wild probability is less than the first threshold value, the test sample is farmed silver carp.

[0024] In an embodiment of the eighth aspect, the step of determining the first threshold value includes: using the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid of the collected muscle samples as a training set; training a binary logistic regression model with the training set to obtain a plurality of the first wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each of the first wild probability values, and drawing an ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each first wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the first wild probability corresponding to the maximum Youden index is the first threshold value. In some embodiments, the first threshold value is 0.50.

[0025] In a ninth aspect, an embodiment discloses a method for distinguishing wild silver carp from farmed silver carp.

[0026] In an embodiment of the ninth aspect, the method comprises a method for testing the content of each biomarker in the biomarkers described in the third aspect in a test sample. In some embodiments, the test sample is an upper back muscle sample of silver carp. In some embodiments, each biomarker comprises proline, valine, and leucine.

[0027] In an embodiment of the ninth aspect, the method further includes: obtaining a second logistic regression model and a second threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the second logistic regression model to obtain a second wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp based on the magnitude of the second wild probability and the second threshold. In some embodiments, the second logistic regression model is: X=-5.711×A+8.571×B+3.255×C-31.835; second wild probability=1 / (1+e －X); wherein A, B, and C are the standardized values ​​of proline, valine, and leucine, respectively, and the second wild probability represents the probability that the test sample is wild silver carp. In some embodiments, if the second wild probability is greater than the second threshold value, the test sample is wild silver carp. In some embodiments, if the second wild probability is less than the second threshold value, the test sample is farmed silver carp.

[0028] In an embodiment of the ninth aspect, the step of determining the second threshold value includes: using the standardized values ​​of proline, valine, and leucine of the collected muscle samples as a training set; training a binary logistic regression model with the training set to obtain a plurality of the second wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each of the second wild probability values, and drawing an ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each second wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the second wild probability corresponding to the maximum Youden index is the second threshold value. In some embodiments, the second threshold value is 0.544.

[0029] In a tenth aspect, an embodiment discloses a method for distinguishing wild silver carp from farmed silver carp.

[0030] In an embodiment of the tenth aspect, the method includes a method for testing the content of each biomarker in the biomarkers described in the fourth aspect in a test sample. In some embodiments, the test sample is an upper back muscle sample of silver carp. In some embodiments, the individual biomarkers include Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0031] In an embodiment of the tenth aspect, the method further includes: obtaining a third logistic regression model and a third threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the third logistic regression model to obtain a third wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp based on the third wild probability and the third threshold. In some embodiments, the third logistic regression model is: X=-673.687×A-367.827×B-570.700×C+568.811×D+345.444×E+5080.475; third wild probability=1 / (1+e －X); wherein A, B, C, D, and E are the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine, respectively, and the third wild probability represents the probability that the test sample is wild silver carp. In some embodiments, if the third wild probability is greater than the third threshold value, the test sample is wild silver carp. In some embodiments, if the third wild probability is less than the third threshold value, the test sample is farmed silver carp.

[0032] In an embodiment of the tenth aspect, the step of determining the third threshold value includes: using the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine of the collected muscle samples as a training set; training a binary logistic regression model with the training set to obtain a plurality of the third wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each of the third wild probability values, and drawing an ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each third wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the third wild probability corresponding to the maximum Youden index is the third threshold value. In some embodiments, the third threshold value is 0.50.

[0033] In an eleventh aspect, an embodiment discloses a method for distinguishing wild silver carp from farmed silver carp.

[0034] In an embodiment of the eleventh aspect, the method comprises a method for testing the content of each biomarker in the biomarkers described in the fifth aspect in a test sample. In some embodiments, the test sample is an upper back muscle sample of silver carp. In some embodiments, the individual biomarkers include DL-2-aminocaprylic acid and 5-aminolevulinic acid.

[0035] In an embodiment of the eleventh aspect, the method further includes: obtaining a fourth logistic regression model and a fourth threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the fourth logistic regression model to obtain a fourth wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp based on the fourth wild probability and the fourth threshold. In some embodiments, the fourth logistic regression model is: X = 4.583 × A-2.879 × B-8.170; fourth wild probability = 1 / (1+e －X); wherein A and B are the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid, respectively, and the fourth wild probability represents the probability that the test sample is wild silver carp. In some embodiments, if the fourth wild probability is greater than the fourth threshold value, the test sample is wild silver carp. In some embodiments, if the fourth wild probability is less than the fourth threshold value, the test sample is farmed silver carp.

[0036] In the embodiment of the eleventh aspect, the step of determining the fourth threshold value includes: using the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid of the collected muscle samples as a training set; training a binary logistic regression model with the training set to obtain a plurality of the fourth wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each of the fourth wild probability values, and drawing an ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each fourth wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the fourth wild probability corresponding to the maximum Youden index is the fourth threshold value. In some embodiments, the fourth threshold value is 0.356.

[0037] In an embodiment of any one of the eighth to eleventh aspects, the standardized value is a value obtained by normalizing the chromatographic peak area detected by each marker and performing a logarithmic transformation with a base of 10.

[0038] In an embodiment of any one of the eighth to eleventh aspects, the method further comprises a liquid chromatography-mass spectrometry method.

[0039] In an embodiment of any one of the eighth to eleventh aspects, in the liquid chromatography detection step, a SHIMADZU-LC30 ultra high performance liquid chromatography system (UHPLC) is used for detection, and an ACQUITY HSS T3 (2.1×100 mm, 1.8 μm) (Waters, Milford, MA, USA) chromatographic column, column temperature 40°C, chromatographic mobile phase A is 0.1% formic acid aqueous solution, chromatographic mobile phase B is 100% acetonitrile solution, elution mode is gradient elution, elution time is 15 minutes, flow rate is 0.3 mL / min, specific injection volume is 4 μL. The specific procedure of chromatographic gradient elution is as follows: 0→2 min, 0% B; 2-6 min, B changes linearly from 0% to 48%; 6→10 min, B changes linearly from 48% to 100%; 10→12 min, B is maintained at 100%; 12→12.1 min, B changes linearly from 100% to 0%, 12.1→15 min, B is maintained at 0%.

[0040] In an embodiment of any one of the eighth to eleventh aspects, in the mass spectrometry detection step, electrospray ionization (ESI) is used for positive ion (+) and negative ion (-) mode detection. After UPLC separation, mass spectrometry analysis is performed using a QE Plus mass spectrometer (ThermoScientific), and ionization is performed using a HESI source, and the ionization conditions are as follows: spray voltage is 3.8 kV (+) and 3.2 kV (-); capillary temperature is 320°C; probe heater temperature is 350°C; sheath gas is 30 arbitrary units; auxiliary gas flow is 5 arbitrary units; S-Lens radio frequency level is 50 arbitrary units.

[0041] In an embodiment of any one of the eighth to eleventh aspects, in the mass spectrometry detection step, the mass spectrometry acquisition time is 15 minutes. The parent ion scanning range is 75-1050 m / z; the primary mass spectrometry resolution is 70,000 @ m / z 200; the number of target ions is 3e6; the maximum injection time of the primary mass spectrometry is 100 ms. The secondary mass spectrometry analysis is collected according to the following method: after each full scan, the secondary mass spectrometry of the 10 highest intensity parent ions is triggered, and the secondary mass spectrometry resolution is 17,500 @ m / z 200; the number of target ions is 1e5, and the maximum injection time of the secondary mass spectrometry is 50 ms.

[0042] In an embodiment of any one of the eighth to eleventh aspects, the results obtained by the liquid chromatography-mass spectrometry are subjected to quality control. The quality control step comprises: preparing a quality control (QC) sample by mixing aliquots of the supernatant from all samples; evaluating the repeatability of the data and the stability of the instrument based on the overlapping base peak ion chromatogram (BPC) of the QC sample; determining the residue using the BPC of the blank sample; and evaluating the reliability and stability of the instrument analysis using principal component analysis (PCA) of all samples.

[0043] In an embodiment of any one of the eighth to eleventh aspects, the raw data obtained by liquid chromatography-mass spectrometry is subjected to peak alignment, retention time correction and peak area extraction using MSDIAL software. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 The metabolic marker composition provided in the embodiment (proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, 5-aminolevulinic acid) is used to distinguish the confusion matrix test results of wild silver carp and farmed silver carp.

[0045] Figure 2 The metabolic marker composition (proline, valine, leucine) provided in the example is used to distinguish the confusion matrix test results of wild silver carp and farmed silver carp.

[0046] Figure 3 The metabolic marker composition (Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine) provided in the examples is used to distinguish the confusion matrix test results of wild silver carp and farmed silver carp.

[0047] Figure 4 The metabolic marker composition (DL-2-aminocaprylic acid, 5-aminolevulinic acid) provided in the example is used to distinguish the confusion matrix test results of wild silver carp and farmed silver carp. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in conjunction with the examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. The reagents not described separately in detail in the present application are all conventional reagents and can be obtained from commercial channels; the methods not described in detail are all conventional experimental methods and can be obtained from the prior art.

[0049] Mining of amino acid biomarkers for distinguishing wild and farmed silver carp

[0050] 1. Collection of muscle samples of silver carp

[0051] The wild silver carp populations involved in the examples come from the Minjiang River, Dongting Lake and Poyang Lake in the Yangtze River Basin. The farmed silver carp populations involved in the examples come from farms in Hubei Province, Hunan Province and Jiangsu Province. The samples were taken from the upper back muscles of the silver carp and immediately frozen in liquid nitrogen after sampling, and transferred to a -80℃ refrigerator for storage until the subsequent metabolite extraction.

[0052] 2. Sample metabolite extraction

[0053] The sample was thawed at 4°C, 50 mg of the sample was weighed, 2 medium-sized steel balls were added, 200 μL of pre-cooled methanol-water solution (4:1, v / v) was added, the sample was homogenized and broken in a tissue disruptor, and 800 μL of pre-cooled 80% methanol-water solution was added, vortexed and mixed, ultrasonicated in an ice bath for 20 min, left to stand overnight at -20°C, centrifuged at 16000g and 4°C for 20 min, the supernatant was taken, and the supernatant was evaporated in a high-speed vacuum concentrator centrifuge. For mass spectrometry detection, 100 μL of methanol-water solution (1:1, v / v) was added for re-dissolution, centrifuged at 20000g and 4°C for 15 min, and the supernatant was taken for UHPLC-MS / MS analysis.

[0054] 3. UHPLC-MS / MS Detection and Analysis

[0055] UHPLC-MS / MS analysis Metabolomics analysis was performed using a UPLC-ESI-Q-Orbitrap-MS system (UHPLC, Shimadzu Nexera X2LC-30AD, Shimadzu Corporation, Japan) combined with Q-Exactive Plus (Thermo Scientific, San Jose, USA). The samples were placed in an autosampler at 4 °C throughout the analysis. The samples were analyzed using a SHIMADZU-LC30 ultra-high performance liquid chromatography system (UHPLC) and ACQUITY HSS T3 (2.1×100 mm, 1.8 μm) (Waters, Milford, MA, USA) chromatographic column, column temperature 40°C, chromatographic mobile phase A is 0.1% formic acid aqueous solution, chromatographic mobile phase B is 100% acetonitrile solution, elution mode is gradient elution, elution time is 15 minutes, flow rate is 0.3 mL / min, specific injection volume is 4 μL. The specific procedure of chromatographic gradient elution is as follows: 0→2 min, 0% B; 2→6 min, B changes linearly from 0% to 48%; 6→10 min, B changes linearly from 48% to 100%; 10→12 min, B is maintained at 100%; 12→12.1 min, B changes linearly from 100% to 0%, 12.1→15 min, B is maintained at 0%.

[0056] Each sample was detected by electrospray ionization (ESI) in positive (+) and negative (-) modes. After UPLC separation, the samples were analyzed by mass spectrometry using a QE Plus mass spectrometer (Thermo Scientific), using a HESI source for ionization. The ionization conditions were as follows: spray voltage of 3.8 kV (+) and 3.2 kV (-); capillary temperature of 320 °C; probe heater temperature of 350 °C; sheath gas of 30 arbitrary units; auxiliary gas flow of 5 arbitrary units; S-Lens radio frequency level of 50 arbitrary units.

[0057] The mass spectrometry acquisition settings were as follows: the mass spectrometry acquisition time was 15 min. The parent ion scan range was 75-1050 m / z; the primary mass spectrometry resolution was 70,000 @ m / z 200; the number of target ions was 3e6; and the maximum primary injection time was 100 ms. The secondary mass spectrometry analysis was acquired according to the following method: the secondary mass spectra of the 10 highest intensity parent ions were triggered after each full scan, the secondary mass spectrometry resolution was 17,500 @ m / z 200; the number of target ions was 1e5, and the maximum secondary injection time was 50 ms.

[0058] 4. Quality Control

[0059] Quality control (QC) samples were prepared by pooling aliquots of supernatant from all samples. Data repeatability and instrument stability were evaluated based on the overlapping base peak ion chromatograms (BPCs) of the QC samples. Residues were determined using the BPCs of blank samples. The reliability and stability of the instrument analysis were evaluated using principal component analysis (PCA) of all samples.

[0060] 5. Processing and analysis of metabolomics data

[0061] The raw data were aligned, retained and corrected using MSDIAL software, and the peak areas were extracted. The metabolite structures were identified by accurate mass matching (mass deviation mass tolerance <10ppm) and secondary spectrum matching (mass deviation mass tolerance <0.01Da), and public databases such as HMDB, MassBank, GNPS, and the self-built Baipu metabolite standard library (BP-DB) were searched. For the extracted data, the ion peaks with missing values ​​>50% within the group were deleted and did not participate in the subsequent statistical analysis; the total peak area of ​​the positive and negative ion data was normalized, the positive and negative ion peaks were integrated, and Python software was used for pattern recognition. The data was preprocessed by Unit variance scaling (UV) before subsequent data analysis. In this study, the raw data files (.wiff format) were first imported into Progenesis QI software for data preprocessing procedures. This included reviewing the steps of alignment, peak selection, and peak detection, and performing data normalization at the same time. The HMDB and LIPID MAPS public databases were used in the annotation of metabolites. Finally, a data matrix containing m / z, retention time (RT), peak intensity, sample name, and metabolite name is generated.

[0062] 6. Screening of significantly different metabolites

[0063] The data matrix was imported into SIMCA 14.1 software for multivariate statistical analysis, including principal component analysis (PCA) and orthogonal projection latent variable structure discriminant analysis (OPLS-DA). After the OPLS-DA model was constructed, the model was validated and evaluated for overfitting by 200 permutation tests. Then, based on the differentially changed metabolites (DMs) with P value < 0.05, Fold change > 1.5 or Fold change < 0.667, and VIP value > 1.0 obtained from the OPLS-DA model, the significantly different metabolites between wild and farmed silver carp were obtained.

[0064] 7. Statistical analysis methods

[0065] GraphPad Prism 8.0 software was used to perform statistical analysis on the biochemical analysis data. Student's t-test was used to calculate the differences in data means between different groups. SPSS was used to construct and analyze the binary logistic regression model, and the input data were the standardized values ​​of each metabolite, that is, the logarithmic transformation value (log10) with a base of 10 after peak area normalization, which was used to reduce the skewed distribution of the overall values ​​and make the data close to a normal distribution. The experimental results are presented in the form of mean plus or minus standard deviation (SD), where the asterisk (*) indicates that the difference between the experimental group and the control group is statistically significant. Specifically, *P<0.05 indicates a significant difference, and **P<0.01 indicates an extremely significant difference.

[0066] 8. Results

[0067] The constructed OPLS-DA model was used to compare the two groups of silver carp samples from wild and farmed populations. Metabolites with OPLS-DA VIP values ​​greater than 1.0 obtained by variance analysis were considered to be potential biomarkers that contribute to class discrimination. The combination of difference fold (FC) ≥1.5 or ≤0.667 and P<0.05 was used to screen significantly differential metabolites based on P value and VIP value. The study of significantly differential metabolites in a single fish species is easily affected by the inherent characteristics and errors of that specific species, while the analysis of significantly differential metabolites shared by different fish species can avoid the errors of a single fish species and improve the accuracy and reliability of the results.

[0068] Therefore, the embodiment obtained the significant difference metabolites shared by these four fish species under wild and culture conditions by performing VEEN intersection processing on the wild and cultured differential metabolites of silver carp. Subsequently, cluster analysis was performed for these shared significant difference metabolites, and it was found that 10 kinds of amino acid substances were included, which were respectively: proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid. The reference mass-to-charge ratio, retention time (min), ionization mode, and subclass of the above-mentioned amino acid substances are shown in Table 1, and the P value, difference multiple, Log2 (difference multiple), and OPLS-DAVIP of the significant difference metabolites of each wild and cultured silver carp are shown in Table 2.

[0069] Table 1 Biomarker information with significant differences between wild and farmed silver carp

[0070]

[0071] Table 2 Differences in biomarkers common to wild and farmed silver carp

[0072]

[0073]

[0074] Metabolic marker composition for distinguishing wild silver carp from cultured silver carp and its use

[0075] The main body of the current data discrimination model is to build a model through multiple indicators for comprehensive judgment. Among them, logistic regression, as a simple and intuitive model, has many advantages. First, it is easy to understand and explain, does not require complex mathematical derivation or calculation, and has high computational efficiency. Secondly, logistic regression can provide highly interpretable result output, can be classified by setting thresholds, and the coefficients and intercepts of the model can provide the degree and direction of the contribution of the variable to the target variable.

[0076] 1. Selection of metabolic marker composition

[0077] The embodiment is based on liquid phase detection of the content of metabolites in muscle samples of wild silver carp and farmed silver carp, and a binary logistic regression model is trained according to the detection results. Compounds with high statistical significance, small measurement error, and high data integrity are selected as biomarkers for distinguishing wild silver carp from farmed silver carp. The biomarkers for distinguishing wild silver carp from farmed silver carp provided in the embodiment include proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid.

[0078] 2. How to distinguish wild silver carp from farmed silver carp

[0079] The embodiment also discloses a method for distinguishing wild silver carp from farmed silver carp. The method comprises: obtaining a first logistic regression model and a first threshold; obtaining multiple standardized values ​​of the contents of multiple biomarkers in a test sample; inputting the multiple standardized values ​​into the first logistic regression model to obtain a first wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp according to the magnitude of the first wild probability and the first threshold. The multiple biomarkers include proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid.

[0080] In some embodiments, the step of obtaining a first logistic regression model includes:

[0081] 1) Obtain training samples

[0082] Embodiment 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp were subjected to the same liquid phase detection as above to obtain the chromatographic peak areas of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid in each sample. These chromatographic peak areas were standardized to obtain standardized values. The standardization step includes: normalizing the chromatographic peak areas of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid, respectively, and performing a logarithmic transformation with a base of 10 to obtain standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid, respectively. The standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid of these muscle samples are used as training samples.

[0083] 2) Training

[0084] The standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid are used as independent variables, and the probability of the sample being wild is used as the dependent variable to train the binary logistic regression model. In some embodiments, the training process of the binary logistic regression model includes: performing composition analysis on the dependent variable and the independent variable to determine whether they meet the prerequisites of the logistic regression; performing a significance test on the independent variable, including degrees of freedom, significance, and scoring; performing a Hosmer-Lemmeshow test to verify the applicability and rationality of the model and ensure the accuracy of the model; training the binary logistic regression model, and obtaining a first logistic regression model according to the degree of influence of the independent variable on the dependent variable.

[0085] In some steps, the embodiment performs a single degree of freedom test on the content of muscle samples of wild silver carp and farmed silver carp with independent variables of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid, including giving degrees of freedom, significance, and score detection. The results are shown in Table 3. The overall statistical score of the content of these four biomarkers as independent variables is high, and the significance is less than 0.05, indicating that it will have a significant impact on the corresponding variables in general.

[0086] Table 3 Significance test and score of independent variables in the logistic regression model of biomarkers of wild and farmed silver carp

[0087]

[0088] In some steps, the embodiment performs a Hosmer-Lemmeshaw test on the logistic regression model of biomarkers of wild silver carp and farmed silver carp. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P=1.000>0.05, that is, the 0 hypothesis is accepted. The results are shown in Table 4. The model fits the real data well and can truly and reliably reflect the reliable relationship between the original variables.

[0089] Table 4 Hosmer-Lemmeshow test for logistic regression models of biomarkers in wild and farmed silver carp

[0090] step Chi-square Degrees of Freedom Significance 1 <0.001 7 1.000

[0091] 3) Obtain the first logistic regression model

[0092] In some steps, the embodiment performs binary logistic regression analysis on the logistic regression model of biomarkers of wild silver carp and farmed silver carp. Table 5 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0093] Table 5 Analysis of binary logistic regression model in biomarker model of wild and farmed silver carp

[0094]

[0095] Embodiment According to the standardized values ​​of the above 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp, the first logistic regression model is as follows:

[0096] X=-23.428×A+38.212×B+28.738×C-63.957×D-124.040×E-57.712×F+50.106×G+47.266×H+15.842×I+9.120×J+665.268;

[0097] First wild probability = 1 / (1+e －X );

[0098] Among them, A, B, C, D, E, F, G, H, I, and J are the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid, respectively. The first wild probability represents the probability that the test sample is wild silver carp.

[0099] The steps provided in the embodiment are calculated by ROC curve to obtain the first threshold. The steps provided in some embodiments include: using the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminocaprylic acid, and 5-aminolevulinic acid of 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp as training sets; training a binary logistic regression model with the training set to obtain multiple first wild probability values ​​as described above; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each first wild probability value, and drawing the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each predicted wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the first predicted wild probability corresponding to the maximum Youden index is the first threshold.

[0100] In some embodiments, if the first wild probability is greater than the first threshold, the test sample is determined to be wild silver carp. If the first wild probability is less than the first threshold, the test sample is determined to be farmed silver carp.

[0101] In some embodiments, the first threshold value obtained by using the above test set and calculating the ROC curve is 0.50. That is, when the first wild probability is greater than 0.50, the test sample is determined to be wild silver carp; when the first wild probability is less than 0.50, the test sample is determined to be farmed silver carp.

[0102] 3. Identification application and testing

[0103] In some test cases, 54 muscle samples (including 30 wild and 24 farmed) were selected, including proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid as biomarkers, and their standardized values ​​were input into the above-mentioned first logistic regression model to obtain the first wild probability; based on the size of the first wild probability and the first threshold value (0.50), it was determined whether the test sample was wild silver carp or farmed silver carp.

[0104] The standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid in the above training set were used to construct binary logistic regression models separately, and the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid were combined to construct a first logistic regression model. The 12 samples were tested respectively, and the discrimination results were subjected to ROC curve test. The results are shown in Table 6. As shown in Table 6, the first logistic regression model constructed by combining the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid had higher AUC values, sensitivity, and specificity.

[0105] Table 6 ROC test results of biomarkers and their combinations for distinguishing wild and farmed silver carp

[0106] Biomarkers AUC Sensitivity Specificity Proline 0.886 0.633 1.000 Valine 0.846 0.700 0.917 Leucine 0.722 0.467 0.917 Nε-Formyl-L-lysine 0.924 0.967 0.875 GABA 0.842 0.867 0.792 N-Formylglycine 0.803 0.800 0.75 N-Acetyl-L-Lysine 0.776 0.867 0.667 Creatinine 0.933 0.967 0.792 DL-2-aminocaprylic acid 0.908 0.867 0.833 5-Aminolevulinic acid 0.901 0.933 0.833 Composition 0.994 0.967 1.000

[0107] The discrimination results were input into the SPSS20.0 analysis software for confusion matrix test. The confusion matrix, also known as the error matrix, is a standard format for expressing accuracy evaluation, expressed in the form of a matrix with n rows and n columns. Specific evaluation indicators include overall accuracy, mapping accuracy, user accuracy, etc. These accuracy indicators reflect the accuracy of image classification from different aspects. Each column of the confusion matrix represents a predicted category, and the total number of each column represents the number of data predicted to be in that category; each row represents the true category of the data, and the total number of data in each row represents the number of data instances in that category. The results are as follows: Figure 1 As shown in the figure, only one sample out of 54 silver carp muscle samples was misjudged, and the overall correct rate of cross-validation was as high as 98.15%. This shows that proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid can be used as biomarkers to identify wild and farmed silver carp in the Yangtze River Basin. The excellent ROC curve results, good correct diagnosis rate, and good complementarity of biomarkers all indicate that the four amino acids have the potential to identify wild and farmed silver carp.

[0108] Metabolic marker composition for distinguishing wild silver carp from cultured silver carp and its use

[0109] 1. Selection of metabolic marker composition

[0110] The embodiment is based on liquid phase detection of the content of metabolites in muscle samples of wild silver carp and farmed silver carp, and training of a binary logistic regression model is performed according to the detection results. Compounds with high statistical significance, small measurement error, and high data integrity are selected as biomarkers for distinguishing wild silver carp from farmed silver carp. The biomarkers for distinguishing wild silver carp from farmed silver carp provided in the embodiment include proline, valine, and leucine.

[0111] The embodiment also provides a kit for detecting these biomarkers, which includes reagents for detecting these biomarkers, such as reagents required for liquid chromatography-mass spectrometry.

[0112] 2. How to distinguish wild silver carp from farmed silver carp

[0113] The embodiment also discloses a method for distinguishing wild silver carp from farmed silver carp. The method includes: obtaining a second logistic regression model and a second threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the second logistic regression model to obtain a second wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp based on the magnitude of the second wild probability and the second threshold. The multiple biomarkers include proline, valine, and leucine.

[0114] In some embodiments, the step of obtaining a second logistic regression model includes:

[0115] 1) Obtain training samples

[0116] Embodiment The muscle samples of 60 wild silver carp and 48 cultured silver carp were tested by the same liquid phase as above to obtain the chromatographic peak areas of proline, valine and leucine in each sample. These chromatographic peak areas were standardized to obtain standardized values. The steps of the standardized treatment include: respectively normalizing the chromatographic peak areas of proline, valine and leucine, and performing logarithmic conversion with a base of 10 to obtain the standardized values ​​of proline, valine and leucine respectively. The standardized values ​​of proline, valine and leucine of these muscle samples were used as training samples.

[0117] 2) Training

[0118] The standardized values ​​of proline, valine, and leucine are used as independent variables, and the probability of the sample being wild is used as the dependent variable to train the binary logistic regression model. In some embodiments, the training process of the binary logistic regression model includes: composition analysis of the dependent variable and the independent variable to determine whether they meet the prerequisites of logistic regression; significance test of the independent variable, including degrees of freedom, significance and score; Hosmer-Lemmeshow test to verify the applicability and rationality of the model and ensure the accuracy of the model; training the binary logistic regression model to obtain a second logistic regression model according to the degree of influence of the independent variable on the dependent variable.

[0119] In some steps, the embodiment performs a single degree of freedom test on the muscle sample content of the independent variables proline, valine, and leucine, including giving degrees of freedom, significance, and score detection. The results are shown in Table 7. The overall statistical score of the content of these four biomarkers as independent variables is high, and the significance is less than 0.001, indicating that the independent variable will have a very significant impact on the corresponding variable in general.

[0120] Table 7 Significance tests and scores of independent variables in the logistic regression model of biomarkers of wild and farmed silver carp

[0121]

[0122]

[0123] In some steps, the embodiment performs a Hosmer-Lemmeshaw test on the logistic regression model of biomarkers of wild silver carp and farmed silver carp. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P=0.710>0.05, that is, the 0 hypothesis is accepted. The results are shown in Table 8. The constructed binary logistic regression model has a good fit with the real data and can truly and reliably reflect the reliable relationship between the original variables.

[0124] Table 8 Hosmer-Lemmeshaw test for logistic regression model of wild and farmed silver carp

[0125] step Chi-square Degrees of Freedom Significance 1 5.438 8 0.710

[0126] 3) Obtain the second logistic regression model

[0127] In some steps, the embodiment performs binary logistic regression analysis on the biomarkers for distinguishing wild silver carp from farmed silver carp. Table 9 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0128] Table 9 Binary logistic regression analysis of wild and farmed silver carp

[0129]

[0130] Embodiment According to the standardized values ​​of the above 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp, the first logistic regression model is as follows:

[0131] X=-5.711×A+8.571×B+3.255×C-31.835;

[0132] Second wild probability = 1 / (1+e －X );

[0133] Among them, A, B, and C are the standardized values ​​of proline, valine, and leucine, respectively, and the second wild probability represents the probability that the test sample is wild silver carp.

[0134] The steps provided in the embodiment are to obtain the second threshold value by ROC curve calculation. The steps provided in some embodiments include: using the standardized values ​​of proline, valine and leucine of 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp as training sets; training a binary logistic regression model with the training set to obtain multiple second wild probability values ​​as described above; respectively calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each second wild probability value, and drawing the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each predicted wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the first predicted wild probability corresponding to the maximum Youden index is the second threshold value.

[0135] In some embodiments, if the second wild probability is greater than the second threshold, the test sample is determined to be wild silver carp. If the second wild probability is less than the second threshold, the test sample is determined to be farmed silver carp.

[0136] In some embodiments, the second threshold value obtained by using the above test set and calculating the ROC curve is 0.544. That is, when the second wild probability is greater than 0.544, the test sample is determined to be wild silver carp; when the second wild probability is less than 0.544, the test sample is determined to be farmed silver carp.

[0137] 3. Identification application and testing

[0138] In some test cases, 54 muscle samples (including 30 wild and 24 farmed) of proline, valine, and leucine were selected as biomarkers, and their standardized values ​​were input into the above-mentioned second logistic regression model to obtain the second wild probability; based on the size of the second wild probability and the second threshold (0.544), it was determined whether the test sample was wild silver carp or farmed silver carp.

[0139] The standardized values ​​of proline, valine, and leucine in the above training set were used to construct binary logistic regression models separately, and the standardized values ​​of proline, valine, and leucine were combined to construct a second logistic regression model. The 24 samples were tested respectively, and the discrimination results were subjected to ROC curve test. The results are shown in Table 10. As can be seen from Table 10, the second logistic regression model constructed by combining the standardized values ​​of proline, valine, and leucine for discrimination of the test samples has a higher AUC value, sensitivity, and specificity.

[0140] Table 10 ROC test results of biomarkers and their combinations for distinguishing wild and farmed silver carp

[0141] Biomarkers AUC Sensitivity Specificity Proline 0.886 0.633 1.000 Valine 0.846 0.700 0.917 Leucine 0.722 0.467 0.917 Composition 0.918 0.867 0.833

[0142] The discrimination results were input into SPSS20.0 analysis software for confusion matrix test. The results are as follows: Figure 2 As shown, only 8 samples among the 54 silver carp samples were misjudged, and the overall correct rate of interactive verification was as high as 85.2%.

[0143] This shows that proline, valine and leucine can be used as biomarkers to identify wild and farmed silver carp in the Yangtze River Basin. The excellent ROC curve results, good correct diagnosis rate and good complementarity of biomarkers all indicate that the four amino acids have the potential to identify wild and farmed silver carp.

[0144] Metabolic marker composition for distinguishing wild silver carp from cultured silver carp and its use

[0145] 1. Selection of metabolic marker composition

[0146] The embodiment is based on liquid phase detection of the content of metabolites in muscle samples of wild silver carp and farmed silver carp, and training of a binary logistic regression model is performed according to the detection results. Compounds with high statistical significance, small measurement error, and high data integrity are selected as biomarkers for distinguishing wild silver carp from farmed silver carp. The biomarkers for distinguishing wild silver carp from farmed silver carp provided in the embodiment include Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0147] The embodiment also provides a kit for detecting these biomarkers, which includes reagents for detecting these biomarkers, such as reagents required for liquid chromatography-mass spectrometry.

[0148] 2. How to distinguish wild silver carp from farmed silver carp

[0149] The embodiment also discloses a method for distinguishing wild silver carp from farmed silver carp. The method includes: obtaining a third logistic regression model and a third threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the third logistic regression model to obtain a third wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp according to the size of the third wild probability and the third threshold. Among them, the multiple biomarkers include Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine.

[0150] In some embodiments, the step of obtaining the third logistic regression model includes:

[0151] 1) Obtain training samples

[0152] Embodiment The muscle samples of 60 wild silver carp and the muscle samples of 48 farmed silver carp were detected by the same liquid phase as above, and the chromatographic peak areas of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine and creatinine in each sample were obtained. These chromatographic peak areas were standardized to obtain standardized values. The step of the standardized treatment includes: after the chromatographic peak areas of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine and creatinine were normalized, the logarithmic conversion with a base of 10 was performed, and the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine and creatinine were obtained respectively. The standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine of these muscle samples were used as training samples.

[0153] 2) Training

[0154] The standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine are used as independent variables, and the probability of the sample being wild is used as the dependent variable to train the binary logistic regression model. In some embodiments, the training process of the binary logistic regression model includes: composition analysis of the dependent variable and the independent variable to determine whether they meet the prerequisites of the logistic regression; significance test of the independent variable, including degrees of freedom, significance and score; Hosmer-Lemmeshow test to verify the applicability and rationality of the model and ensure the accuracy of the model; training the binary logistic regression model to obtain a third logistic regression model according to the degree of influence of the independent variable on the dependent variable.

[0155] In some steps, the embodiment performs a single degree of freedom test on muscle samples of independent variables Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine, including providing degrees of freedom, significance, and score detection. As shown in Table 11, the overall statistical score of the independent variable is high, and the significance is less than 0.0001, indicating that the independent variable will have a very significant impact on the corresponding variable in general.

[0156] Table 11 Significance test and score of independent variables in the logistic regression model of biomarkers of wild and farmed silver carp

[0157]

[0158] In order to further analyze the influence of constants and independent variables on dependent variables in terms of multiple degrees of freedom, the embodiment also performs a Hosmer-Lemmeshaw test on the logistic regression model. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P=1.000>0.05, that is, the 0 hypothesis is accepted. As shown in the result table 12, the constructed binary logistic regression model has a good fit with the real data and can truly and reliably reflect the reliable relationship between the original variables.

[0159] Table 12 Hosmer-Lemmeshaw test for logistic regression model of wild and farmed silver carp

[0160] step Chi-square Degrees of Freedom Significance 1 <0.001 5 1.000

[0161] 3) Obtain the third logistic regression model

[0162] In some steps, the embodiment performs binary logistic regression analysis on the logistic regression model of biomarkers of wild silver carp and farmed silver carp. Table 13 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0163] Table 13 Binary logistic regression analysis of wild and farmed silver carp

[0164]

[0165] Embodiment According to the standardized values ​​of the above 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp, the third logistic regression model is as follows:

[0166] X=-673.687×A-367.827×B-570.700×C+568.811×D+345.444×E+5080.475;

[0167] The third wild probability = 1 / (1+e －X );

[0168] Among them, A, B, C, D, and E are the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine, respectively, and the third wild probability represents the probability that the test sample is wild silver carp.

[0169] The steps provided in the embodiments are to obtain the third threshold value by ROC curve calculation. The steps provided in some embodiments include: using the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine of 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp as training sets; training a binary logistic regression model with the training set to obtain multiple third wild probability values ​​as described above; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each third wild probability value, and plotting the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each predicted wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the third predicted wild probability corresponding to the maximum Youden index is the third threshold value.

[0170] In some embodiments, if the third wild probability is greater than the third threshold, the test sample is determined to be wild silver carp. If the third wild probability is less than the third threshold, the test sample is determined to be farmed silver carp.

[0171] In some embodiments, the third threshold value obtained by using the above test set and calculating the ROC curve is 0.50. That is, when the third wild probability is greater than 0.50, the test sample is determined to be wild silver carp; when the third wild probability is less than 0.50, the test sample is determined to be farmed silver carp.

[0172] 3. Identification application and testing

[0173] In some test cases, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine of 54 muscle samples (including 30 wild and 24 farmed) were selected as biomarkers, and their standardized values ​​were input into the above-mentioned third logistic regression model to obtain the third wild probability; based on the size of the third wild probability and the third threshold value (0.50), it was determined whether the test sample was wild silver carp or farmed silver carp.

[0174] The standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine in the above training set were used to construct binary logistic regression models separately, and the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine were combined to construct a third logistic regression model. The 12 samples were tested respectively, and the discrimination results were subjected to ROC curve test. The results are shown in Table 14. As can be seen from Table 14, the third logistic regression model constructed by combining the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine has a higher AUC value, sensitivity, and specificity.

[0175] Table 14 ROC test results of biomarkers and their combinations for distinguishing wild and farmed silver carp

[0176] Biomarkers AUC Sensitivity Specificity Nε-Formyl-L-lysine 0.924 0.967 0.875 GABA 0.842 0.867 0.792 N-Formylglycine 0.803 0.800 0.750 N-Acetyl-L-Lysine 0.776 0.867 0.667 Creatinine 0.933 0.967 0.792 Combined markers 0.989 0.967 0.917

[0177] The discrimination results were input into SPSS20.0 analysis software for confusion matrix test. The results are as follows: Figure 3 As shown in the figure, only 3 samples out of 54 silver carp samples were misjudged, and the overall accuracy of the interactive validation was as high as 94.4%.

[0178] This shows that the logistic regression model constructed with Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine as biomarkers and their standardized values ​​can be used to distinguish wild and farmed silver carp in the Yangtze River Basin. The excellent ROC curve results and good correct diagnosis rate as well as the good complementarity of biomarkers indicate that the five amino acids have the potential to identify wild and farmed silver carp.

[0179] Metabolic marker composition for distinguishing wild silver carp from cultured silver carp and its use

[0180] 1. Selection of metabolic marker composition

[0181] The embodiment is based on liquid phase detection of the content of metabolites in muscle samples of wild silver carp and farmed silver carp, and training of a binary logistic regression model is performed according to the detection results. Compounds with high statistical significance, small measurement error, and high data integrity are selected as biomarkers for distinguishing wild silver carp from farmed silver carp. The biomarkers for distinguishing wild silver carp from farmed silver carp provided in the embodiment include DL-2-aminocaprylic acid and 5-aminolevulinic acid.

[0182] The embodiment also provides a kit for detecting these biomarkers, which includes reagents for detecting these biomarkers, such as reagents required for liquid chromatography-mass spectrometry.

[0183] 2. How to distinguish wild silver carp from farmed silver carp

[0184] The embodiment also discloses a method for distinguishing wild silver carp from farmed silver carp. The method includes: obtaining a fourth logistic regression model and a fourth threshold; obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample; inputting the multiple standardized values ​​into the fourth logistic regression model to obtain a fourth wild probability; and determining whether the test sample is the wild silver carp or the farmed silver carp according to the size of the fourth wild probability and the fourth threshold. Among them, the multiple biomarkers include DL-2-aminocaprylic acid and 5-aminolevulinic acid.

[0185] In some embodiments, the step of obtaining the fourth logistic regression model includes:

[0186] 1) Obtain training samples

[0187] Embodiment The muscle samples of 60 wild silver carps and the muscle samples of 48 cultured silver carps are detected by the same liquid phase, and the chromatographic peak areas of DL-2-aminocaprylic acid and 5-aminolevulinic acid in each sample are obtained. These chromatographic peak areas are subjected to standardization treatment to obtain standardized values. The step of this standardized treatment comprises: after the chromatographic peak areas of DL-2-aminocaprylic acid and 5-aminolevulinic acid are subjected to normalization treatment respectively, the logarithmic conversion with the base number of 10 carried out can obtain the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid respectively. The standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid of these muscle samples are as training samples.

[0188] 2) Training

[0189] The standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid are used as independent variables, and the probability of the sample being wild is used as the dependent variable to train the binary logistic regression model. In some embodiments, the training process of the binary logistic regression model includes: composition analysis of the dependent variable and the independent variable to determine whether they meet the prerequisites of the logistic regression; significance test of the independent variable, including degrees of freedom, significance and score; Hosmer-Lemmeshow test to verify the applicability and rationality of the model and ensure the accuracy of the model; training the binary logistic regression model to obtain a fourth logistic regression model according to the degree of influence of the independent variable on the dependent variable.

[0190] In some steps, the embodiment performs a single degree of freedom test on the muscle sample content of the independent variables DL-2-aminocaprylic acid and 5-aminolevulinic acid, including giving degrees of freedom, significance and score detection. The results are shown in Table 15, and the overall statistical scores of these independent variables are high, and the significance is less than 0.0001, indicating that the independent variables have a very significant impact on the corresponding variables in general.

[0191] Table 15 Significance test and score of independent variables in the joint biomarker model of wild and farmed silver carp

[0192]

[0193] In some steps, the embodiment performs a Hosmer-Lemmeshaw test on the logistic regression model of biomarkers of wild silver carp and farmed silver carp. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P=0.072>0.05, that is, the 0 hypothesis is accepted. As shown in Table 16, the constructed binary logistic regression model fits the real data well and can truly and reliably reflect the reliable relationship between the original variables.

[0194] Table 16 Hosmer-Lemmeshaw test for binary logistic regression model of wild and farmed silver carp

[0195] step Chi-square Degrees of Freedom Significance 1 23.667 8 0.072

[0196] 3) Obtain the fourth logistic regression model

[0197] In some steps, the embodiment performs binary logistic regression analysis on the biomarkers for distinguishing wild silver carp from farmed silver carp. Table 17 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0198] Table 17 Binary logistic regression analysis of wild and farmed silver carp

[0199]

[0200] Embodiment According to the standardized values ​​of the above 60 wild silver carp muscle samples and 48 farmed silver carp muscle samples, the fourth logistic regression model is as follows:

[0201] X=4.583×A-2.879×B-8.170;

[0202] Fourth wild probability = 1 / (1+e －X );

[0203] A and B are the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid, respectively, and the fourth wild probability represents the probability that the test sample is wild silver carp.

[0204] The steps provided in the embodiment are to obtain the fourth threshold value by ROC curve calculation. The steps provided in some embodiments include: using the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid of 60 muscle samples of wild silver carp and 48 muscle samples of farmed silver carp as training sets; training a binary logistic regression model with the training set to obtain multiple fourth wild probability values ​​as described above; respectively calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each fourth wild probability value, and drawing the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index corresponding to each predicted wild probability (Youden index = sensitivity + specificity - 1), and obtaining the maximum Youden index, the fourth predicted wild probability corresponding to the maximum Youden index is the fourth threshold value.

[0205] In some embodiments, if the fourth wild probability is greater than the fourth threshold, the test sample is determined to be wild silver carp. If the fourth wild probability is less than the fourth threshold, the test sample is determined to be farmed silver carp.

[0206] In some embodiments, the fourth threshold value obtained by using the above test set and calculating the ROC curve is 0.356. That is, when the fourth wild probability is greater than 0.356, the test sample is determined to be wild silver carp; when the fourth wild probability is less than 0.356, the test sample is determined to be farmed silver carp.

[0207] 3. Identification application and testing

[0208] In some test cases, 54 muscle samples (including 30 wild and 24 farmed) DL-2-aminocaprylic acid and 5-aminolevulinic acid were selected as biomarkers, and their standardized values ​​were input into the fourth logistic regression model to obtain the fourth wild probability; based on the size of the fourth wild probability and the fourth threshold (0.356), it was determined whether the test sample was wild silver carp or farmed silver carp.

[0209] The standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid in the above training set were used to construct binary logistic regression models separately, and the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid were combined to construct a fourth logistic regression model. The 54 muscle samples were tested respectively, and the discrimination results were subjected to ROC curve test, and the results are shown in Table 18. As can be seen from Table 18, the fourth logistic regression model constructed by combining the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid for discrimination of test samples has a higher AUC value, sensitivity and specificity.

[0210] Table 18 ROC test results of biomarkers and their combinations for distinguishing wild and farmed silver carp

[0211] Biomarkers AUC Sensitivity Specificity DL-2-aminocaprylic acid 0.908 0.867 0.833 5-Aminolevulinic acid 0.901 0.933 0.833 Logo composition 0.942 1.000 0.917

[0212] The discrimination results were input into SPSS20.0 analysis software for confusion matrix test, such as Figure 4 As shown in the figure, only 2 samples among the 54 silver carp samples were misjudged, and the overall correct rate of the interactive validation was 96.3%.

[0213] This shows that the stability of using DL-2-aminocaprylic acid and 5-aminolevulinic acid as biomarkers and regression models for distinguishing wild silver carp from farmed silver carp is good. The excellent ROC curve results, good correct diagnosis rate and good complementarity of biomarkers indicate that the five amino acids have the potential to distinguish wild silver carp from farmed silver carp.

[0214] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.

Claims

1. Use of a biomarker and / or a reagent for detecting the biomarker in distinguishing wild silver carp from farmed silver carp, the distinction comprising: Testing the content of biomarkers in the test sample, wherein the biomarkers include proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid; obtaining a first logistic regression model and a first threshold; Obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample, wherein the standardized values ​​are logarithmic transformation values ​​with a base of 10 after peak area normalization; Inputting the plurality of the standardized values ​​into the first logistic regression model to obtain a first wild probability; Determine whether the test sample is the wild silver carp or the farmed silver carp according to the first wild probability and the first threshold; If the first wild probability is greater than the first threshold, the test sample is a wild silver carp; as well as If the first wild probability is less than the first threshold, the test sample is farmed silver carp; Among them, the first logistic regression model is: X=-23.428×A+38.212×B+28.738×C-63.957×D-124.040×E-57.712×F+50.106×G+47.266×H+15.842×I+9.120×J+665.268; the first wild probability=1 / (1+e-X); wherein A, B, C, D, E, F, G, H, I, and J are the standardized values ​​of proline, valine, leucine, Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, creatinine, DL-2-aminooctanoic acid, and 5-aminolevulinic acid, respectively, and the first wild probability represents the probability that the test sample is wild silver carp.

2. Use of biomarkers and / or reagents for detecting the biomarkers in distinguishing wild silver carp from farmed silver carp, the distinction comprising: Testing the content of biomarkers in the test sample, wherein the biomarkers include proline, valine, and leucine; Obtain a second logistic regression model and a second threshold; Obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample, wherein the standardized values ​​are logarithmic transformation values ​​with a base of 10 after peak area normalization; Inputting the plurality of the standardized values ​​into the second logistic regression model to obtain a second wild probability; Determine whether the test sample is the wild silver carp or the farmed silver carp according to the second wild probability and the second threshold; If the second wild probability is greater than the second threshold, the test sample is a wild silver carp; as well as If the second wild probability is less than the second threshold, the tested sample is farmed silver carp; Among them, the second logistic regression model is: X=-5.711×A+8.571×B+3.255×C-31.835; the second wild probability=1 / (1+e-X); wherein A, B, and C are the standardized values ​​of proline, valine, and leucine, respectively, and the second wild probability represents the probability that the test sample is wild silver carp.

3. Use of biomarkers and / or reagents for detecting the biomarkers in distinguishing wild silver carp from farmed silver carp, the distinction comprising: Testing the content of biomarkers in the test sample, wherein the biomarkers include Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine; Obtaining a third logistic regression model and a third threshold; Obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample, wherein the standardized values ​​are logarithmic transformation values ​​with a base of 10 after peak area normalization; Inputting the plurality of standardized values ​​into the third logistic regression model to obtain a third wild probability; Determine whether the test sample is the wild silver carp or the farmed silver carp according to the third wild probability and the third threshold; If the third wild probability is greater than the third threshold, the test sample is a wild silver carp; and If the third wild probability is less than the third threshold, the tested sample is farmed silver carp; Among them, the third logistic regression model is: X=-673.687×A-367.827×B-570.700×C+568.811×D+345.444×E+5080.475; the third wild probability=1 / (1+e-X); wherein A, B, C, D, and E are the standardized values ​​of Nε-formyl-L-lysine, γ-aminobutyric acid, N-formylglycine, N-acetyl-L-lysine, and creatinine, respectively, and the third wild probability represents the probability that the test sample is wild silver carp.

4. Use of biomarkers and / or reagents for detecting the biomarkers in distinguishing wild silver carp from farmed silver carp, the distinction comprising: Testing the content of biomarkers in the test sample, wherein each biomarker includes DL-2-aminocaprylic acid and 5-aminolevulinic acid; obtaining a fourth logistic regression model and a fourth threshold; Obtaining multiple standardized values ​​of the content of multiple biomarkers in the test sample, wherein the standardized values ​​are logarithmic transformation values ​​with a base of 10 after peak area normalization; Inputting the plurality of standardized values ​​into the fourth logistic regression model to obtain a fourth wild probability; Determining whether the test sample is the wild silver carp or the farmed silver carp according to the fourth wild probability and the fourth threshold; If the fourth wild probability is greater than the fourth threshold, the test sample is a wild silver carp; as well as If the fourth wild probability is less than the fourth threshold, the tested sample is farmed silver carp; The fourth logistic regression model is: X = 4.583 × A-2.879 × B-8.170; the fourth wild probability = 1 / (1+e －X ); wherein A and B are the standardized values ​​of DL-2-aminocaprylic acid and 5-aminolevulinic acid, respectively, and the fourth wild probability characterizes the probability that the test sample is a wild silver carp.

Images