Discriminating fatty acid biomarkers for wild and cultured hypophthalmichthys nobilis, methods and uses thereof

Abstract

The application relates to the technical field of judging or confirming hypophthalmichthys nobilis, and particularly relates to a fatty acid biomarker for distinguishing wild and cultured hypophthalmichthys nobilis, a method and an application. The biomarker comprises at least one of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid and cis-5, 8, 11, 14, 17-eicosapentaenoic acid. The biomarker can accurately distinguish wild and cultured hypophthalmichthys nobilis, provide scientific basis and technical support for relevant law enforcement departments, strengthen supervision and strike force of illegal fishing behavior in the Yangtze River basin, reduce overfishing of wild resources, and further provide a technical basis for better management and protection of hypophthalmichthys nobilis resources.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of determining or confirming Hypophthalmichthys nobilis, in particular to a fatty acid biomarker for distinguishing wild and cultured Hypophthalmichthys nobilis, a method and an application. BACKGROUND

[0002] Hypophthalmichthys nobilis is an important cultured species in China, mainly distributed in the Yangtze River Basin. Due to the influence of water pollution, overfishing and other factors, the fish biodiversity in the Yangtze River Basin is decreasing, so China has implemented the "Yangtze River Ten-Year Fishing Ban" policy since 2020, and has cracked down on illegal fishing and market sales of Yangtze River catch in the Yangtze River Basin. However, due to the difficulty in distinguishing between wild and cultured aquatic products in the Yangtze River Basin at present, it is difficult to meet the requirements of the "Yangtze River Fishing Ban" policy. Therefore, developing a determination or confirmation technology for wild and cultured Hypophthalmichthys nobilis, mining biological markers of wild and cultured Hypophthalmichthys nobilis, and promoting its application are reliable ways to support the "Yangtze River Fishing Ban" policy. SUMMARY

[0003] In view of the close biological characteristics of wild and cultured Hypophthalmichthys nobilis and the lack of distinguishing technology, the present application first uses a gas chromatograph to collect muscle samples of different wild and cultured Hypophthalmichthys nobilis populations, performs fatty acid spectrum analysis, and explores the common differences and changes in the fatty acids of wild and cultured Hypophthalmichthys nobilis, and selects fatty acid biomarkers: tetradecanoic acid (CAS: 57677-52-8, fatty acid abbreviation: C14:0), cis, cis, cis-9, 12, 15-octadecatrienoic acid (CAS: 463-40-1, fatty acid abbreviation: C18:3n3), cis-5, 8, 11, 14-eicosatetraenoic acid (CAS: 463-40-1, fatty acid abbreviation: C20:4n6), cis-5, 8, 11, 14, 17-eicosapentaenoic acid (CAS: 463-40-1, fatty acid abbreviation: C20:5n3). The results of gas chromatograph detection are subjected to statistical and logistic regression analysis to obtain a logistic regression analysis model, which can accurately distinguish the source of the test sample and has good sensitivity and specificity. Moreover, the fatty acid biomarkers and the application provided by the present application can provide scientific basis and technical support for relevant law enforcement departments, strengthen the supervision and crackdown on illegal fishing behavior in the Yangtze River Basin, reduce overfishing of wild resources, and thus provide a technical basis for better management and protection of Hypophthalmichthys nobilis resources.

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

[0005] In a first aspect, embodiments disclose biomarkers containing at least one of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid.

[0006] In a second aspect, embodiments disclose a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a first logistic regression model and a first threshold value; obtaining a plurality of relative content percentages of a plurality of biomarkers in a test sample, the plurality of biomarkers being tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid; inputting the plurality of relative content percentages into the first logistic regression model to obtain a first wild probability, the first wild probability representing the probability that the test sample is wild Hypophthalmichthys nobilis; and determining whether the test sample is the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis according to the size of the first wild probability and the first threshold value. In some embodiments, the first logistic regression model is X = 1.435 × A - 2.169 × B - 1.198 × C + 0.164 × D + 17.337; the first wild probability = 1 / (1+e －X ); wherein A, B, C, D are the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid, respectively.

[0007] In some embodiments of the second aspect, if the first wild probability is greater than the first threshold value, the test sample is wild Hypophthalmichthys nobilis. If the first wild probability is less than the first threshold value, the test sample is cultured Hypophthalmichthys nobilis.

[0008] In embodiments of the second aspect, the determination of the first threshold value comprises: using the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic 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 first wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each first wild probability value, respectively, and plotting a ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each first wild probability, and obtaining the maximum Youden index, wherein the first threshold value is the first wild probability corresponding to the maximum Youden index. In some embodiments, the first threshold value is 0.4742.

[0009] In a third aspect, the embodiments disclose a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises obtaining a second logistic regression model and a second threshold value; obtaining a plurality of relative content percentages of a plurality of biomarkers in a test sample, the plurality of biomarkers being tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, and cis-5,8,11,14-eicosatetraenoic acid; inputting the plurality of relative content percentages into the second logistic regression model to obtain a second wild probability, the second wild probability representing a probability that the test sample is wild Hypophthalmichthys nobilis; and determining, according to the size of the second wild probability and the second threshold value, whether the test sample is the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis. In some embodiments of the third aspect, the second logistic regression model is X = 1.444 × A - 2.179 × B - 1.071 × C + 18.038, and the second wild probability = 1 / (1+e －X ); wherein A, B, and C are the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, and cis-5,8,11,14-eicosatetraenoic acid, respectively.

[0010] In some embodiments of the third aspect, if the second wild probability is greater than the second threshold value, the test sample is wild Hypophthalmichthys nobilis. If the second wild probability is less than the second threshold value, the test sample is cultured Hypophthalmichthys nobilis.

[0011] In some embodiments of the third aspect, the determination of the second threshold value comprises: using the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, and cis-5,8,11,14-eicosatetraenoic 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 second wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each second wild probability value, respectively, and plotting a ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each second wild probability, and obtaining the maximum Youden index, wherein the second wild probability corresponding to the maximum Youden index is the second threshold value. In some embodiments, the second threshold value is 0.4677.

[0012] In a fourth aspect, the embodiments disclose a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a third logistic regression model and a third threshold value; obtaining a plurality of relative content percentages of a plurality of biomarkers in a test sample, the plurality of biomarkers being cis,cis,cis-9,12,15-octadecatrienoic acid and cis-5,8,11,14-eicosatetraenoic acid; inputting the plurality of relative content percentages into the third logistic regression model to obtain a third wild probability, the third wild probability representing a probability that the test sample is wild Hypophthalmichthys nobilis; and determining, according to the size of the third wild probability and the third threshold value, whether the test sample is the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis. In some embodiments of the fourth aspect, the third logistic regression model is X = -1.781 × A - 1.505 × B + 21.011, and the third wild probability = 1 / (1+e －X ); wherein A and B are the relative content percentages of cis,cis,cis-9,12,15-octadecatrienoic acid and cis-5,8,11,14-eicosatetraenoic acid, respectively.

[0013] In some embodiments of the fourth aspect, if the third wild probability is greater than the third threshold value, the test sample is wild Hypophthalmichthys nobilis. If the third wild probability is less than the third threshold value, the test sample is cultured Hypophthalmichthys nobilis.

[0014] In some embodiments of the fourth aspect, the determination of the third threshold value comprises: using the relative content percentages of cis,cis,cis-9,12,15-octadecatrienoic acid and cis-5,8,11,14-eicosatetraenoic 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 third wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each third wild probability value, respectively, and plotting a ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each third wild probability, and obtaining the maximum Youden index, wherein the third wild probability corresponding to the maximum Youden index is the third threshold value. In some embodiments, the third threshold value is 0.7032.

[0015] In a fifth aspect, embodiments disclose a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a fourth logistic regression model and a fourth threshold value; obtaining a plurality of relative content percentages of a plurality of biomarkers in a test sample, the plurality of biomarkers being tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid; inputting the plurality of relative content percentages into the fourth logistic regression model to obtain a fourth wild probability, the fourth wild probability representing the probability that the test sample is wild Hypophthalmichthys nobilis; and determining whether the test sample is the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis according to the size of the fourth wild probability and the fourth threshold value. The fourth logistic regression model is X = 2.724 × A - 2.675 × B + 11.124, and the fourth wild probability = 1 / (1+e －X ); wherein A and B are the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid, respectively.

[0016] In some embodiments of the fifth aspect, if the fourth wild probability is greater than the fourth threshold value, the test sample is wild Hypophthalmichthys nobilis. If the fourth wild probability is less than the fourth threshold value, the test sample is cultured Hypophthalmichthys nobilis.

[0017] In some embodiments of the fifth aspect, the determination of the fourth threshold value comprises: taking the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic 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 fourth wild probability values; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each fourth wild probability value, respectively, and plotting a ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each fourth wild probability, and obtaining the maximum Youden index, wherein the fourth wild probability corresponding to the maximum Youden index is the fourth threshold value. In some embodiments, the fourth threshold value is 0.4672.

[0018] In a sixth aspect, embodiments disclose the use of the biomarkers of the first aspect and / or the reagents for detecting the biomarkers of the first aspect in distinguishing wild and cultured Hypophthalmichthys nobilis. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 Confusion matrix test results of the fatty acid biomarker composition (tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid, cis-5, 8, 11, 14, 17-eicosapentaenoic acid) provided by embodiments for distinguishing wild and cultured Hypophthalmichthys nobilis.

[0020] Figure 2Confusion matrix test results of the fatty acid biomarker composition (tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid) provided by the examples to distinguish wild silver carp from farmed silver carp.

[0021] Figure 3 Confusion matrix test results of the fatty acid biomarker composition (cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid) provided by the examples to distinguish wild silver carp from farmed silver carp.

[0022] Figure 4 Confusion matrix test results of the fatty acid biomarker composition (tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid) provided by the examples to distinguish wild silver carp from farmed silver carp. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application. The reagents not specifically described in the present application are conventional reagents and can be obtained from commercial channels; the methods not specifically described are conventional experimental methods and can be known from prior art.

[0024] Mining of fatty acid biomarkers to discriminate wild and cultured hypophthalmichthys nobilis

[0025] 1. Muscle sample collection of silver carp

[0026] The wild silver carp populations involved in the examples are respectively from the Minjiang River, the Jialing River, the Three Gorges Reservoir, the Dongting Lake and the Poyang Lake of the Yangtze River. The farmed silver carp populations involved in the examples are respectively from the farms in Hubei Province, Hunan Province and Jiangsu Province. The samples are all taken from the upper back muscle of silver carp, and are immediately frozen and stored in liquid nitrogen after sampling, and are transferred to a-80℃ refrigerator for storage, and are stored for subsequent fatty acid extraction.

[0027] 2. Determination of fatty acid content of samples

[0028] The determination of fatty acid content refers to the internal standard method of GB 5009.168-2016 "Determination of fatty acids in food". This method can determine the content of 37 kinds of fatty acids. The specific process is as follows: accurately weigh about 1.00 g of muscle dry sample into a hydrolysis tube, add 1 mL of 10000 mg / L glyceryl tricapride internal standard, hydrolyze with 10 mL of hydrochloric acid, then extract with 25 mL of mixed solution of equal volume of ether and petroleum ether, and take the supernatant into a chicken heart bottle. Repeat the operation three times. Concentrate the extract to dryness by rotary evaporation, add 8 mL of 2% sodium hydroxide methanol solution for saponification, then add 7 mL of 14% boron trifluoride methanol solution for methyl esterification, finally add 10 mL of n-hexane, mix well by vortex, and take 1 mL of supernatant to the sample bottle after standing and layering. Determine using Agilent 7890A gas chromatograph.

[0029] Chromatographic column: CP-Sil 88 capillary column (100 m x 0.25 mm, 0.20 μm); temperature program: initial temperature 100℃, hold for 13 min, increase to 180℃ at a rate of 10℃ / min, hold for 6 min, then increase to 200℃ at a rate of 1℃ / min, hold for 20 min, finally increase to 230℃ at a rate of 4℃ / min, hold for 15 min. The temperatures of the injection port and the detector are set to 270℃ and 280℃ respectively.

[0030] After gas chromatographic analysis, each peak in the chromatogram obtained represents a specific fatty acid methyl ester. By using chromatographic data processing software, the area of each peak can be calculated. By comparing the peak area of the target fatty acid methyl ester with the peak area of the internal standard, the absolute content of the fatty acid can be calculated. After obtaining the absolute content of each fatty acid, the relative percentage of each fatty acid in the total fatty acids can be calculated by the following formula: Relative percentage of a specific fatty acid = (content of the specific fatty acid) / (total content of all fatty acids) x 100%.

[0031] 3. Statistical analysis method

[0032] The t-test was used to calculate the significant differences in fatty acid data between different wild and cultured populations of Hypophthalmichthys nobilis. The relative percentage of each fatty acid was entered into SPSS for the construction and analysis of binary logistic regression model. The relative percentage of each fatty acid was presented in the form of mean plus or minus standard deviation (S.D.), and the asterisk (*) indicated that the difference between the experimental group and the control group was statistically significant. Specifically, *P<0.05 indicated significant difference, and **P<0.01 indicated extremely significant difference.

[0033] 4. Results

[0034] The combination of fold change ≥ 1.5 or ≤ 0.667 and P < 0.05 is used to screen the significantly different fatty acids by comparing the two groups of samples of wild and cultured Hypophthalmichthys nobilis. Four significantly different fatty acids are found through analysis, which are, in turn, tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid, and cis-5, 8, 11, 14, 17-eicosapentaenoic acid. The relative content percentage, fold change, and P value of the above fatty acid substances are shown in Table 1. In Table 1, the fold change is the ratio of the relative content percentage of wild Hypophthalmichthys nobilis to that of cultured Hypophthalmichthys nobilis.

[0035] Table 1 Difference in fold change of fatty acid biomarkers between wild and cultured Hypophthalmichthys nobilis

[0036]

[0037] Fatty acid biomarker composition to discriminate wild and cultured hypophthalmichthys nobilis and uses thereof

[0038] The current data discrimination model is constructed by multiple indicators. Among them, logistic regression, as a simple and intuitive model, has many advantages. First, it is easy to understand and explain, without complex mathematical derivation or calculation, and it is high in computing efficiency. Second, logistic regression can provide strong interpretability of the results output, which can be classified by setting a threshold, and the coefficients and intercepts of the model can provide the contribution degree and direction of the variables to the target variable.

[0039] 1. Selection of fatty acid biomarker composition

[0040] The embodiment is based on the detection of the content of fatty acids in the muscle samples of wild Hypophthalmichthys nobilis and cultured Hypophthalmichthys nobilis, and the binary logistic regression model is trained according to the detection results. The fatty acids with high statistical significance, small measurement error, and high data integrity are selected as biomarkers for discriminating wild Hypophthalmichthys nobilis and cultured Hypophthalmichthys nobilis. The fatty acid biomarkers provided by the embodiment for discriminating wild Hypophthalmichthys nobilis and cultured Hypophthalmichthys nobilis include tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid, and cis-5, 8, 11, 14, 17-eicosapentaenoic acid.

[0041] 2. Method for discriminating wild Hypophthalmichthys nobilis and cultured Hypophthalmichthys nobilis

[0042] The embodiments also disclose a method for distinguishing wild Amur catfish from farmed Amur catfish. The method comprises: obtaining a first logistic regression model and a first threshold value; obtaining a plurality of relative content percentages of a plurality of biomarkers in a test sample; inputting the plurality of relative content percentages into the first logistic regression model to obtain a first wild probability; and determining whether the test sample is the wild Amur catfish or the farmed Amur catfish according to the size relationship between the first wild probability and the first threshold value. The plurality of biomarkers comprises tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, and cis-5,8,11,14,17-eicosapentaenoic acid.

[0043] In some embodiments, the step of obtaining the first logistic regression model comprises:

[0044] 1) obtaining training samples

[0045] In the embodiments, muscle samples of 30 wild Amur catfish and muscle samples of 30 farmed Amur catfish are detected by using the same gas chromatography as described above to obtain the peak areas of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, and cis-5,8,11,14,17-eicosapentaenoic acid in each sample. The absolute contents of target fatty acid and total fatty acid are calculated by comparing the peak area of the target fatty acid methyl ester with the peak area of the internal standard, and the relative content percentage of the target fatty acid is obtained according to the formula: (the content of a specific fatty acid / the total content of all fatty acids) × 100. The relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, and cis-5,8,11,14,17-eicosapentaenoic acid in these muscle samples are taken as training samples.

[0046] 2) training

[0047] In the embodiments, the training of the binary logistic regression model comprises: performing component analysis on the dependent variable and the independent variable to determine whether they meet the prerequisites of logistic regression; performing significance test on the independent variable, including degrees of freedom, significance, and score; performing Hosmer-Lemeshow test to verify the applicability and rationality of the model and ensure the accuracy of the model; and training the binary logistic regression model to obtain the first logistic regression model according to the influence of the independent variable on the dependent variable.

[0048] In some embodiments, the example performs a one degree of freedom test on the content of the independent variables tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid, cis-5, 8, 11, 14, 17-eicosapentaenoic acid in muscle samples of wild and cultured grass carp, including giving degrees of freedom, significance and score test. The results are shown in Table 2, the overall statistical score of the content of the four biomarkers as independent variables is high, the significance is less than 0.05, indicating that it will have a significant impact on the independent variable as a whole.

[0049] Table 2 Significance test and score of independent variables of the logistic regression model of biomarkers of wild and cultured grass carp

[0050]

[0051] In some embodiments, the example performs a Hosmer-Lemeshow test on the logistic regression model of biomarkers of wild and cultured grass carp. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P = 1.000 > 0.05, i.e. accept the 0 hypothesis. The results are shown in Table 3, the model is well fitted with the true data, and can truly and reliably reflect the reliable relationship between the original variables.

[0052] Table 3 Hosmer-Lemeshow test of the logistic regression model of biomarkers of wild and cultured grass carp

[0053] Step Chi-square Degrees of freedom Significance 1 0.277 7 1.000

[0054] 3) Obtain the first logistic regression model

[0055] In some embodiments, the example performs a binary logistic regression analysis on the logistic regression model of biomarkers of wild and cultured grass carp. Table 4 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0056] Table 4 Binary logistic regression analysis in the biomarker model of wild and cultured grass carp

[0057]

[0058] The first logistic regression model obtained by the example according to the relative content percentage of the 30 muscle samples of wild grass carp and the 30 muscle samples of cultured grass carp is as follows:

[0059] X = 1.435 × A - 2.169 × B - 1.198 × C + 0.164 × D + 17.337;

[0060] The first wild probability = 1 / (1+e －X );

[0061] wherein A, B, C, D are the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid, respectively, and the first wild probability represents the probability that the test sample is wild grass carp.

[0062] The step provided by the embodiments is to calculate the first threshold value by the ROC curve. The steps provided by some embodiments include: taking the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid of 30 muscle samples of wild grass carp and 30 muscle samples of cultured grass carp as a training set; training a binary logistic regression model with the training set to obtain a plurality of 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, respectively, 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, wherein the first predicted wild probability corresponding to the maximum Youden index is the first threshold value.

[0063] In some embodiments, if the first wild probability is greater than the first threshold value, the test sample is determined to be wild grass carp. If the first wild probability is less than the first threshold value, the test sample is determined to be cultured grass carp.

[0064] In some embodiments, the first threshold value obtained by calculating the ROC curve with the above test set is 0.4742. That is, when the first wild probability is greater than 0.4742, the test sample is determined to be wild grass carp; when the first wild probability is less than 0.4742, the test sample is determined to be cultured grass carp.

[0065] 3. Discrimination application and test

[0066] In some test examples, the relative content percentages of tetradecanoic acid, cis,cis,cis-9,12,15-octadecatrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, cis-5,8,11,14,17-eicosapentaenoic acid of 55 muscle samples (including 29 wild and 26 cultured) are selected as biomarkers, and the relative content percentages are input into the above first logistic regression model to obtain the first wild probability; according to the size of the first wild probability and the first threshold value (0.4742), it is determined whether the test sample is wild grass carp or cultured grass carp.

[0067] The relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid and cis-5, 8, 11, 14, 17-eicosapentaenoic acid were respectively used to construct a binary logistic regression model, and a first logistic regression model was constructed by combining the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid and cis-5, 8, 11, 14, 17-eicosapentaenoic acid. The 55 muscle samples were detected, and the discrimination results were subjected to ROC curve test, and the results are shown in Table 5. As shown in Table 5, the first logistic regression model constructed by combining the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid and cis-5, 8, 11, 14, 17-eicosapentaenoic acid has higher AUC value, sensitivity and specificity.

[0068] Table 5 ROC test results of each biomarker and its combination for discriminating wild and cultured hypophthalmichthys nobilis

[0069] Biomarker AUC Sensitivity Specificity Tetradecanoic acid 0.667 1.000 0.308 Cis,cis,cis-9,12,15-octadecatrienoic acid 0.942 0.931 0.846 Cis-5,8,11,14-eicosatetraenoic acid 0.886 0.793 0.846 Cis-5,8,11,14,17-eicosapentaenoic acid 0.772 0.690 0.962 Composition 0.998 0.966 1.000

[0070] The discrimination results were input into SPSS 20.0 analysis software for confusion matrix test. The confusion matrix, also known as error matrix, is a standard format for accuracy evaluation, which is represented in the form of n rows and n columns. The specific evaluation indexes include overall accuracy, mapping accuracy, user accuracy, etc., which reflect the accuracy of image classification from different aspects. Each column of the confusion matrix represents the predicted class, and the total number of each column represents the number of data predicted as the class. Each row represents the true belonging class of the data, and the total number of data in each row represents the number of data instances of the class. As shown in Table 6, only one sample of the 55 samples was misjudged, and the overall accuracy of cross-validation was as high as 98.18%. Figure 1

[0071] Therefore, tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, cis-5, 8, 11, 14-eicosatetraenoic acid and cis-5, 8, 11, 14, 17-eicosapentaenoic acid can be used as biomarkers to identify wild and cultured hypophthalmichthys nobilis in the Yangtze River basin. The excellent ROC curve results, good correct diagnosis rate and good complementarity of the biomarkers all indicate that the four fatty acids have the potential to discriminate wild and cultured hypophthalmichthys nobilis. Fatty acid biomarker composition to discriminate wild and cultured hypophthalmichthys nobilis and uses thereof Step

[0072] 1. Selection of fatty acid biomarker combination

[0073] ​The embodiment is based on detection of the content of fatty acids in muscle samples of wild and cultured Hypophthalmichthys nobilis, and training of a binary logistic regression model according to the detection results. The fatty acids with high statistical significance, small determination error and high data integrity are selected as biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis. The fatty acid biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis provided by the embodiment include tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid.

[0074] The embodiment also provides a kit for detecting the biomarkers, which comprises reagents for detecting the biomarkers, such as reagents required for gas chromatograph inspection.

[0075] 2. A method for distinguishing wild and cultured Hypophthalmichthys nobilis

[0076] The embodiment also discloses a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a second logistic regression model and a second threshold value; obtaining a plurality of relative content percentages of a plurality of biomarker contents in a test sample; inputting the plurality of relative content percentages into the second logistic regression model to obtain a second wild probability; and determining the test sample as the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis according to the size of the second wild probability and the second threshold value. The plurality of biomarkers include tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid.

[0077] In some embodiments, the step of obtaining the second logistic regression model comprises:

[0078] 1) obtaining training samples

[0079] In the embodiment, the muscle samples of 30 wild Hypophthalmichthys nobilis and 30 cultured Hypophthalmichthys nobilis are detected by using the same gas chromatograph as described above to obtain the peak areas of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid in each sample. By comparing the peak area of the target fatty acid methyl ester with the peak area of the internal standard, the absolute content of the target fatty acid and the total fatty acid is calculated, and the relative content percentage of the target fatty acid is obtained according to the formula: (content of specific fatty acid / total content of all fatty acids) x 100. The relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid of these muscle samples are used as training samples.

[0080] 2) training

[0081] In some embodiments, the training process of the binary logistic regression model comprises: performing component analysis on the dependent variable and the independent variables to determine whether they meet the prerequisites of logistic regression; performing significance test on the independent variables, including degree of freedom, significance, and score; performing Hosmer-Lemeshow test to verify the applicability and rationality of the model, and to ensure the accuracy of the model; and performing training of the binary logistic regression model to obtain the second logistic regression model according to the influence of the independent variables on the dependent variable.

[0082] In some steps, embodiments perform single degree of freedom test on the muscle sample content of the independent variables tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid, including degree of freedom, significance, and score detection. The results are shown in Table 7, and the overall statistical score of the content of the three biomarkers as independent variables is high, with a significance of less than 0.001, indicating that the independent variables will have a very significant effect on the dependent variables as a whole.

[0083] Table 6 Significance test and score of independent variables in the logistic regression model of biomarkers of wild and cultured Hypophthalmichthys nobilis

[0084]

[0085] In some steps, embodiments perform Hosmer-Lemeshow test on the logistic regression model of biomarkers of wild and cultured Hypophthalmichthys nobilis. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P = 1.000 > 0.05, i.e., the null hypothesis is accepted. As shown in Table 7, the binary logistic regression model built is well fitted with the true data, and can truly and reliably reflect the reliable relationship between the original variables.

[0086] Table 7 Hosmer-Lemeshow test of the logistic regression model of wild and cultured Hypophthalmichthys nobilis

[0087] Chi-square Degrees of freedom Significance Biomarker 1 0.286 7 1.000

[0088] 3) Obtain the second logistic regression model

[0089] In some steps, embodiments perform binary logistic regression analysis on the biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis. Table 8 shows the effect quantity Exp(B) of the regression model and the significance of each biomarker.

[0090] Table 8 Binary logistic regression analysis of wild and cultured Hypophthalmichthys nobilis

[0091]

[0092] The first logistic regression model is obtained according to the relative content percentages of the muscle samples of the 30 wild Amur ide and the muscle samples of the 30 cultured Amur ide in the above embodiment as follows:

[0093] X = 1.444 * A - 2.179 * B - 1.071 * C + 18.038;

[0094] The second wild probability = 1 / (1 + e －X );

[0095] wherein A, B and C are the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid, respectively, and the second wild probability represents the probability that the test sample is a wild Amur ide.

[0096] The second threshold value is obtained by the ROC curve calculation according to the steps provided in the embodiment. The steps provided in some embodiments include: taking the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid in the muscle samples of the 30 wild Amur ide and the muscle samples of the 30 cultured Amur ide as a training set; training a binary logistic regression model with the training set to obtain a plurality of second wild probability values as described above; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each second wild probability value, respectively, and plotting the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each predicted wild probability, and obtaining the maximum Youden index, wherein the first predicted wild probability corresponding to the maximum Youden index is the second threshold value.

[0097] In some embodiments, if the second wild probability is greater than the second threshold value, the test sample is determined to be a wild Amur ide. If the second wild probability is less than the second threshold value, the test sample is determined to be a cultured Amur ide.

[0098] In some embodiments, the second threshold value obtained by the ROC curve calculation of the test set described above is 0.4677. That is, when the second wild probability is greater than 0.4677, the test sample is determined to be a wild Amur ide; when the second wild probability is less than 0.4677, the test sample is determined to be a cultured Amur ide.

[0099] 3. Discrimination application and test

[0100] In some test examples, the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid of 55 muscle samples (including 29 wild and 26 cultured) are selected as biomarkers, and the second wild probability is obtained by inputting the relative content percentages into the above-mentioned second logistic regression model; according to the size of the second wild probability and the second threshold value (0.4677), it is determined that the test sample is wild or cultured.

[0101] The relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid in the above-mentioned training set are respectively constructed into a binary logistic regression model, and a second logistic regression model is constructed by combining the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid, and the 55 muscle samples are respectively detected, and the discrimination results are tested by ROC curve, and the results are shown in Table 9. As shown in Table 9, the discrimination of the test sample by the second logistic regression model constructed by combining the relative content percentages of tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid has higher AUC value, sensitivity and specificity.

[0102] Table 9 ROC test results of each biomarker and its combination for discriminating wild and cultured hypophthalmichthys nobilis

[0103] AUC Sensitivity Specificity Tetradecanoic acid Cis,cis,cis-9,12,15-octadecatrienoic acid 0.667 1.000 0.308 Cis-5,8,11,14-eicosatetraenoic acid 0.942 0.931 0.846 Composition 0.886 0.793 0.846 Figure 2 0.998 0.966 1.000

[0104] The discrimination results are input into SPSS20.0 analysis software for confusion matrix test, and the results are shown in Table 10. Fatty acid biomarker composition to discriminate wild and cultured hypophthalmichthys nobilis and uses thereof As shown in Table 10, only one sample of the 55 samples is misjudged, and the overall correct rate of cross-validation is as high as 98.18%.

[0105] Therefore, tetradecanoic acid, cis, cis, cis-9, 12, 15-octadecatrienoic acid, and cis-5, 8, 11, 14-eicosatetraenoic acid as biomarkers can be used to identify wild and cultured hypophthalmichthys nobilis in the Yangtze River basin. The excellent ROC curve results, good correct diagnosis rate and good complementarity of the biomarkers all indicate that the three fatty acids have the potential to discriminate wild and cultured hypophthalmichthys nobilis.

[0106] Step

[0107] 1. Selection of fatty acid biomarker combination

[0108] The embodiment is based on detection of the content of fatty acids in muscle samples of wild and cultured Hypophthalmichthys nobilis, and training of a binary logistic regression model according to the detection results. The fatty acids with high statistical significance, small determination error and high data integrity are selected as biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis, and the fatty acid biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis provided by the embodiment include cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid.

[0109] The embodiment also provides a kit for detecting the biomarkers, which comprises reagents for detecting the biomarkers, such as reagents required for gas chromatograph inspection.

[0110] 2. A method for distinguishing wild and cultured Hypophthalmichthys nobilis

[0111] The embodiment also discloses a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a third logistic regression model and a third threshold value; obtaining a plurality of relative content percentages of a plurality of biomarker contents in a test sample; inputting the plurality of relative content percentages into the third logistic regression model to obtain a third wild probability; and determining the test sample as the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis according to the size of the third wild probability and the third threshold value. The plurality of biomarkers include cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid.

[0112] In some embodiments, the step of obtaining the third logistic regression model comprises:

[0113] 1) Obtaining training samples

[0114] In the embodiment, the muscle samples of 30 wild Hypophthalmichthys nobilis and 30 cultured Hypophthalmichthys nobilis are detected by using the same gas chromatograph as described above to obtain the peak areas of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid in each sample. The absolute content of the target fatty acid and the total fatty acid is calculated by comparing the peak area of the target fatty acid methyl ester with the peak area of the internal standard, and the relative content percentage of the target fatty acid is obtained according to the formula: (content of specific fatty acid / total content of all fatty acids) x 100%. The relative content percentages of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid of these muscle samples are used as training samples.

[0115] 2) Training

[0116] In some embodiments, the binary logistic regression model is trained by taking the relative percentage of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid as independent variables, and the probability of the sample being wild as dependent variable. In some embodiments, the training process of the binary logistic regression model comprises: performing component analysis on the dependent variable and the independent variables to determine whether they meet the prerequisites of logistic regression; performing significance test on the independent variables, including degree of freedom, significance, and score; performing Hosmer-Lemeshow test to verify the applicability and rationality of the model, and to ensure the accuracy of the model; and training the binary logistic regression model to obtain a third logistic regression model according to the influence of the independent variables on the dependent variable.

[0117] In some steps, embodiments perform single degree of freedom test on the muscle samples of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid as independent variables, including giving degree of freedom, significance, and score test. As shown in Table 10, the overall statistical score of the independent variables is high, and the significance is less than 0.001, indicating that the independent variables will have a very significant effect on the dependent variable as a whole.

[0118] Table 10 Significance test and score of independent variables in the logistic regression model of biomarkers of wild and cultured Hypophthalmichthys nobilis

[0119]

[0120] To further analyze the influence of the constant and the independent variables on the dependent variable in multiple degrees of freedom, embodiments also perform Hosmer-Lemeshow 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, i.e., the null hypothesis is accepted. As shown in Table 11, the binary logistic regression model built has a good fitting condition with the true data, and can truly and reliably reflect the reliable relationship between the original variables.

[0121] Table 11 Hosmer-Lemeshow test of the logistic regression model of wild and cultured Hypophthalmichthys nobilis

[0122] Chi-square Degrees of freedom Significance Biomarker 1 0.409 7 1.000

[0123] 3) Obtain a third logistic regression model

[0124] In some steps, embodiments perform binary logistic regression analysis on the logistic regression model of biomarkers of wild and cultured Hypophthalmichthys nobilis. Table 12 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0125] Table 12 Binary logistic regression analysis of wild and cultured Hypophthalmichthys nobilis

[0126]

[0127] The third logistic regression model is obtained according to the relative content percentages of the muscle samples of the 30 wild silver carp and the muscle samples of the 30 farmed silver carp in the above embodiment.

[0128] X = -1.781 * A - 1.505 * B + 21.011.

[0129] The third wild probability = 1 / (1 + e －X ).

[0130] Wherein A and B are the relative content percentages of cis-9,12,15-octadecatrienoic acid and cis-5,8,11,14-eicosatetraenoic acid, respectively, and the third wild probability represents the probability that the test sample is a wild silver carp.

[0131] The third threshold is obtained by the ROC curve calculation according to the steps provided in the embodiment. The steps provided in some embodiments include: taking the relative content percentages of cis-9,12,15-octadecatrienoic acid and cis-5,8,11,14-eicosatetraenoic acid in the muscle samples of the 30 wild silver carp and the muscle samples of the 30 farmed silver carp as a training set; training a binary logistic regression model with the training set to obtain a plurality of 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, respectively, and drawing an ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each predicted wild probability, and obtaining the maximum Youden index, wherein the third predicted wild probability corresponding to the maximum Youden index is the third threshold.

[0132] In some embodiments, if the third wild probability is greater than the third threshold, the test sample is determined to be a wild silver carp. If the third wild probability is less than the third threshold, the test sample is determined to be a farmed silver carp.

[0133] In some embodiments, the third threshold obtained by the ROC curve calculation of the test set described above is 0.7032. That is, when the third wild probability is greater than 0.7032, the test sample is determined to be a wild silver carp; when the third wild probability is less than 0.7032, the test sample is determined to be a farmed silver carp.

[0134] 3. Discrimination application and test

[0135] In some test examples, cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid of 55 muscle samples (including 29 wild and 26 cultured) were selected as biomarkers, and the relative content percentage was input into the third logistic regression model to obtain the third wild probability; according to the size of the third wild probability and the third threshold value (0.7032), it was determined that the test sample was wild or cultured.

[0136] The relative content percentage of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid in the above training set was respectively constructed into a binary logistic regression model, and a third logistic regression model was constructed by combining the relative content percentage of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid. The 55 muscle samples were detected, and the discrimination results were tested by ROC curve, as shown in Table 13. As shown in Table 13, the third logistic regression model constructed by combining the relative content percentage of cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid has higher AUC value, sensitivity and specificity.

[0137] Table 13 ROC test results of each biomarker and its combination for discriminating wild and cultured grass carp

[0138] AUC Sensitivity Specificity Cis,cis,cis-9,12,15-octadecatrienoic acid Cis-5,8,11,14-eicosatetraenoic acid 0.942 0.931 0.846 Composition 0.886 0.793 0.846 Figure 3 0.973 0.931 0.962

[0139] The discrimination results were input into SPSS20.0 analysis software for confusion matrix test, and the results are shown in Table 14. Fatty acid biomarker composition to discriminate wild and cultured hypophthalmichthys nobilis and uses thereof As shown in Table 14, only 3 samples of 55 grass carp samples were misjudged, and the overall correct rate of cross-validation was as high as 94.54%.

[0140] Therefore, the logistic regression model constructed by taking cis, cis, cis-9, 12, 15-octadecatrienoic acid and cis-5, 8, 11, 14-eicosatetraenoic acid as biomarkers and using their relative content percentages can be used to identify wild and cultured grass carp in the Yangtze River basin. Excellent ROC curve results, good correct diagnosis rate and good complementarity of biomarkers indicate that the two kinds of fatty acids have the potential to determine wild and cultured grass carp.

[0141] Step

[0142] 1. Selection of fatty acid biomarker combination

[0143] The embodiment is based on detection of the content of fatty acids in muscle samples of wild and cultured Hypophthalmichthys nobilis, and training of a binary logistic regression model according to the detection results. The fatty acids with high statistical significance, small determination error and high data integrity are selected as biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis, and the fatty acid biomarkers for distinguishing wild and cultured Hypophthalmichthys nobilis provided by the embodiment include tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid.

[0144] The embodiment also provides a kit for detecting the biomarkers, which comprises reagents for detecting the biomarkers, such as reagents required for gas chromatograph inspection.

[0145] 2. A method for distinguishing wild and cultured Hypophthalmichthys nobilis

[0146] The embodiment also discloses a method for distinguishing wild and cultured Hypophthalmichthys nobilis. The method comprises: obtaining a fourth logistic regression model and a fourth threshold value; obtaining a plurality of relative content percentages of a plurality of biomarker contents in a test sample; inputting the plurality of relative content percentages into the fourth logistic regression model to obtain a fourth wild probability; and determining the test sample as the wild Hypophthalmichthys nobilis or the cultured Hypophthalmichthys nobilis according to the size of the fourth wild probability and the fourth threshold value. The plurality of biomarkers include tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid.

[0147] In some embodiments, the step of obtaining the fourth logistic regression model comprises:

[0148] 1) Obtaining training samples

[0149] In the embodiment, the muscle samples of 30 wild Hypophthalmichthys nobilis and the muscle samples of 30 cultured Hypophthalmichthys nobilis are detected by using the same gas chromatograph as described above to obtain the peak areas of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid in each sample. By comparing the peak area of the target fatty acid methyl ester with the peak area of the internal standard, the absolute content of the target fatty acid and the total fatty acid is calculated, and the relative content percentage of the target fatty acid is obtained according to the formula: (the content of a specific fatty acid / the total content of all fatty acids) x 100. The relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid of these muscle samples are used as training samples.

[0150] 2) Training

[0151] In some embodiments, the training process of the binary logistic regression model comprises: performing component analysis on the dependent variable and the independent variables to determine whether they meet the prerequisites of logistic regression; performing significance test on the independent variables, including degree of freedom, significance, and score; performing Hosmer-Lemeshow test to verify the applicability and rationality of the model, and to ensure the accuracy of the model; and performing training of the binary logistic regression model to obtain the fourth logistic regression model according to the influence of the independent variables on the dependent variable.

[0152] In some steps, embodiments perform single degree of freedom test on the muscle sample content of the independent variables, tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid, including degree of freedom, significance, and score detection. As shown in Table 14, the overall statistical score of these independent variables is high, and the significance is less than 0.001, indicating that the independent variables will have a very significant effect on the dependent variable as a whole.

[0153] Table 14 Significance test and score of independent variables in joint biomarker model of wild and cultured Hypophthalmichthys nobilis

[0154]

[0155] In some steps, embodiments perform Hosmer-Lemeshow test on the logistic regression model of the biomarkers of wild and cultured Hypophthalmichthys nobilis. Under the condition of multiple degrees of freedom, the significance of the binary logistic regression model is P = 1.000 > 0.05, i.e., accept the null hypothesis. As shown in Table 15, the binary logistic regression model built has a good fitting condition with the true data, and can truly and reliably reflect the reliable relationship between the original variables.

[0156] Table 15 Hosmer-Lemeshow test of binary logistic regression model of wild and cultured Hypophthalmichthys nobilis

[0157] Chi-square Degrees of freedom Significance Biomarker 1 0.424 7 1.000

[0158] 3) Obtain the fourth logistic regression model

[0159] In some steps, embodiments perform binary logistic regression analysis on the biomarkers for discriminating wild and cultured Hypophthalmichthys nobilis. Table 16 shows the effect size Exp(B) of the regression model and the significance of each biomarker.

[0160] Table 16 Binary logistic regression analysis of wild and cultured Hypophthalmichthys nobilis

[0161]

[0162] The fourth logistic regression model is obtained according to the relative content percentages of the muscle samples of the 30 wild Amur bream and the muscle samples of the 30 cultured Amur bream.

[0163] X = 2.724 * A - 2.675 * B + 11.124;

[0164] The fourth wild probability = 1 / (1 + e －X ).

[0165] Wherein A and B are the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid, respectively, and the fourth wild probability represents the probability that the test sample is a wild Amur bream.

[0166] The fourth threshold value is obtained by calculating the ROC curve according to the steps provided in the embodiments. The steps provided in some embodiments include: taking the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid in the muscle samples of the 30 wild Amur bream and the muscle samples of the 30 cultured Amur bream as a training set; training a binary logistic regression model with the training set to obtain a plurality of fourth wild probability values as described above; calculating the sensitivity (true positive rate) and 1-specificity (false positive rate) of each fourth wild probability value, respectively, and plotting the ROC curve with 1-specificity as the abscissa and sensitivity as the ordinate; calculating the Youden index (Youden index = sensitivity + specificity - 1) corresponding to each predicted wild probability, and obtaining the maximum Youden index, wherein the fourth predicted wild probability corresponding to the maximum Youden index is the fourth threshold value.

[0167] In some embodiments, if the fourth wild probability is greater than the fourth threshold value, the test sample is determined to be a wild Amur bream. If the fourth wild probability is less than the fourth threshold value, the test sample is determined to be a cultured Amur bream.

[0168] In some embodiments, the fourth threshold value obtained by calculating the ROC curve with the test set described above is 0.4672. That is, when the fourth wild probability is greater than 0.4672, the test sample is determined to be a wild Amur bream; when the fourth wild probability is less than 0.4672, the test sample is determined to be a cultured Amur bream.

[0169] 3. Discrimination application and test

[0170] In some test examples, the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid in 55 muscle samples (including 29 wild and 26 cultured) are selected as biomarkers, and the fourth wild probability is obtained by inputting the relative content percentages into the fourth logistic regression model described above. According to the size of the fourth wild probability and the fourth threshold value (0.4672), it is determined whether the test sample is a wild Amur bream or a cultured Amur bream.

[0171] The relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid in the above training set were respectively constructed into binary logistic regression models, and the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid were combined to construct a fourth logistic regression model. The 55 muscle samples were detected respectively, and the discrimination results were tested by ROC curve. The results are shown in Table 17. As shown in Table 17, the fourth logistic regression model constructed by combining the relative content percentages of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid has higher AUC value, sensitivity and specificity in discriminating the test samples.

[0172] Table 17 ROC test results of each biomarker and its combination in discriminating wild and cultured Hypophthalmichthys nobilis

[0173] AUC Sensitivity Specificity Tetradecanoic acid Cis,cis,cis-9,12,15-octadecatrienoic acid 0.667 1.000 0.308 Composition 0.942 0.931 0.846 Figure 4 0.985 1.000 0.923

[0174] The discrimination results were input into SPSS20.0 analysis software for confusion matrix test. As shown in Table 18, only 2 samples were misjudged in the 55 Hypophthalmichthys nobilis samples, and the overall accuracy of cross-validation was as high as 96.36%. ​

[0175] Therefore, it is proved that the stability of tetradecanoic acid and cis, cis, cis-9, 12, 15-octadecatrienoic acid as biomarkers and regression model in discriminating wild and cultured Hypophthalmichthys nobilis is good. The excellent ROC curve results, good correct diagnosis rate and good complementarity of biomarkers indicate that the two fatty acids have the potential to discriminate wild and cultured Hypophthalmichthys nobilis.

[0176] The above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.​

Claims

1. Methods for distinguishing between wild and farmed silver carp include: Obtain the first logistic regression model and the first threshold; Obtain multiple relative percentages of the content of multiple biomarkers in the test sample; the multiple biomarkers are tetradecanoic acid, cis,cis,cis-9,12,15-octadecanotrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, and cis-5,8,11,14,17-eicosatepentanoic acid. The first wild probability is obtained by inputting multiple relative content percentages into the first logistic regression model. The first wild probability represents the probability that the test sample is a wild silver carp. The test sample is determined to be either the wild silver carp or the farmed silver carp based on the magnitude of the first wild probability and the first threshold. The first logistic regression model is as follows: X=1.435×A-2.169×B-1.198×C+0.164×D+17.337; The probability of first wild animal = 1 / (1+e) －X ); A, B, C, and D represent the relative percentage contents of tetradecanoic acid, cis, cis, cis-9,12,15-octadecanetrienoic acid, cis-5,8,11,14-eicosatetraenoic acid, and cis-5,8,11,14,17-eicosatepentanoic acid, respectively.

2. The method according to claim 1, wherein if the first wild probability is greater than the first threshold, the test sample is a wild silver carp; If the first wild probability is less than the first threshold, then the test sample is a farmed silver carp.

3. Methods for distinguishing between wild and farmed silver carp include: Obtain the second logistic regression model and the second threshold; Obtain multiple relative percentages of the content of multiple biomarkers in the test sample; the multiple biomarkers are tetradecanoic acid, cis,cis,cis-9,12,15-octadecanotrienoic acid, and cis-5,8,11,14-eicosatetraenoic acid. The second wild probability is obtained by inputting multiple relative content percentages into the second logistic regression model, whereby the second wild probability represents the probability that the test sample is a wild silver carp. The test sample is determined to be either the wild silver carp or the farmed silver carp based on the magnitude of the second wild probability and the second threshold. The second logistic regression model is as follows: X=1.444×A-2.179×B-1.071×C+18.038; The probability of a second wild animal is 1 / (1+e) －X ); A, B, and C represent the relative percentage contents of tetradecanoic acid, cis, cis, cis-9,12,15-octadecanetrienoic acid, and cis-5,8,11,14-eicosatetraenoic acid, respectively.

4. The method according to claim 3, wherein if the second wild probability is greater than the second threshold, the test sample is a wild silver carp; If the second wild probability is less than the second threshold, then the test sample is a farmed silver carp.

5. Methods for distinguishing between wild and farmed silver carp include: Obtain the third logistic regression model and the third threshold; The relative percentages of multiple biomarkers in the test sample were obtained. The multiple biomarkers were cis, cis, cis-9,12,15-octadecanotrienoic acid and cis-5,8,11,14-eicosatetraenoic acid. The third wild probability is obtained by inputting multiple relative content percentages into the third logistic regression model, and the third wild probability represents the probability that the test sample is a wild silver carp; The test sample is determined to be either the wild silver carp or the farmed silver carp based on the magnitude of the third wild probability and the third threshold. The third logistic regression model is as follows: X=-1.781×A-1.505×B+21.011; The probability of a third wild animal is 1 / (1+e) －X ); Where A and B represent the relative percentage contents of cis, cis, cis-9,12,15-octadecadienoic acid and cis-5,8,11,14-eicosatetraenoic acid, respectively.

6. The method according to claim 5, wherein if the third wild probability is greater than the third threshold, the test sample is a wild silver carp; If the third wild probability is less than the third threshold, then the test sample is a farmed silver carp.

7. Methods for distinguishing between wild and farmed silver carp include: Obtain the fourth logistic regression model and the fourth threshold; The relative percentages of multiple biomarkers in the test sample were obtained. The multiple biomarkers were tetradecanoic acid and cis,cis,cis-9,12,15-octadecanoic acid. The fourth wild probability is obtained by inputting multiple relative content percentages into the fourth logistic regression model, and the fourth wild probability represents the probability that the test sample is a wild silver carp; The test sample is determined to be either the wild silver carp or the farmed silver carp based on the magnitude of the fourth wild probability and the fourth threshold. The fourth logistic regression model is as follows: X = 2.724 × A - 2.675 × B + 11.124; The probability of the fourth wild animal is 1 / (1+e) －X ); Where A and B are the relative percentage contents of tetradecanoic acid, cis, cis-9,12,15-octadecanetrienoic acid, respectively.

8. The method according to claim 7, wherein if the fourth wild probability is greater than the fourth threshold, the test sample is a wild silver carp; If the fourth wild probability is less than the fourth threshold, then the test sample is a farmed silver carp.

Images