A method for constructing a correction model for livestock feed composition based on macrobarcoding

Through DNA macro-barcoding technology and correction model, the problem of difficulty in ensuring accuracy in herbivorous feeding composition determination is solved, and accurate evaluation and stable results of complex forage composition are achieved.

CN112599200BActive Publication Date: 2024-11-29CHINA AGRI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011595272.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-29
Publication Date
2024-11-29
Estimated Expiration
2040-12-29

AI Technical Summary

Technical Problem

The existing methods are difficult to ensure accuracy when determining the feeding composition of herbivores, especially in the case of diversified forage composition and low alkane concentrations, and seasonal and regional factors affect the accuracy of the results.

Method used

A correction model for livestock feeding based on DNA macrobarcode technology was used to improve the accuracy of assay by extracting DNA from feces, using ITS2 barcode PCR amplification and Illumina MiSeq sequencing, and combining bioinformatics analysis.

Benefits of technology

The accuracy of the measurement of feeding components is significantly improved, especially in complex composition situations. Through the application of the correction model, the reliability and stability of the measurement results are ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112599200B_ABST
    Figure CN112599200B_ABST
Patent Text Reader

Abstract

The present invention provides a method for constructing a livestock feeding composition correction model based on macrobarcodes, characterized in that it comprises: (1) selecting simulated diets derived from different forages as correction samples, extracting DNA, performing ITS2 barcode PCR amplification, and performing Illumina MiSeq double-end sequencing of the amplicon to obtain raw sequencing data; (2) filtering and quality control processing the raw sequencing data, and performing OTU clustering on the valid sequences; (3) comparing the OTU representative sequences with the NCBI database Blastn to complete species annotation; (4) using macrobarcode technology to measure the forage gene sequence of the sample, obtaining the relative proportion measurement value of the forage DNA sequence, and analyzing to obtain the components of the simulated diet; and (5) using a linear regression method to associate the measurement values ​​of each component of the simulated diet and the true value of the dry matter ratio of each component to establish a correction model for the diet composition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of molecular biology and animal husbandry. Specifically, the present application provides a method for constructing a macrobarcode-based livestock feed composition correction model. Background Art

[0002] The production performance and nutritional status of grazing livestock are closely related to their forage intake. Under grazing conditions, it is extremely important to accurately grasp the composition and amount of livestock feed, which can clarify the nutritional status of livestock and predict livestock production performance, and then provide effective data support and theoretical guidance for making grassland management decisions and optimizing forage resource allocation. The forage resources on natural grasslands are rich and varied. The plant species ingested by herbivorous livestock are very diverse, and they have certain preferences for specific forage species. Accurately measuring the feed composition of herbivorous livestock still faces many difficulties and challenges.

[0003] There are many methods commonly used to estimate the composition of herbivorous animals at home and abroad, mainly including simulated feeding method, pasture feeding method, esophageal fistula method, rumen content method, fecal microanalysis method, plant wax layer indicator method and near-infrared spectroscopy. Saturated alkane technology is currently considered to be the most accurate and objective method. It uses saturated alkanes as endogenous indicators, determines the types and contents of alkanes in each pasture eaten by livestock, and establishes a regression equation with the types and contents of corresponding alkanes in feces, so as to estimate the proportion of animals eating each pasture. This method can accurately estimate the composition of animals with simple diet components (4 to 8 types). However, when the alkane concentration in the pasture is low or the alkane pattern is similar, the estimation accuracy is greatly reduced as the livestock food composition becomes more complicated. In addition, the characteristic pattern and recovery rate of the wax layer indicator will also vary due to factors such as season and region, which directly affects the accuracy of seasonal feeding estimation results.

[0004] DNA barcoding technology refers to a new molecular identification technology that uses a standard, relatively short DNA fragment within the genome to identify species or their variant types. This technology gets rid of the obstacle of traditional identification methods relying on long-term experience. By establishing a standard database, it can achieve rapid and accurate identification of species, which is an innovation in molecular identification methodology. However, DNA barcoding technology relies on traditional Sanger sequencing technology, and it is not yet possible to efficiently and quickly classify and evaluate mixed samples of multiple individuals and species, and it is difficult to obtain complete, high-quality DNA templates from outdated materials. DNA macrobarcoding technology based on high-throughput sequencing technology (also known as second-generation sequencing technology) is expected to overcome this problem.

[0005] DNA macrobarcoding technology, also known as DNA composite barcoding technology, refers to the use of high-throughput technology to simultaneously obtain barcode gene amplicon sequences of many species, and use bioinformatics analysis to identify the taxonomic unit composition of the community. Its working principle is: (1) Extract DNA from environmental samples; (2) Select appropriate DNA barcodes as molecular markers; (3) Use barcode universal primers to obtain amplicons from environmental samples; (4) Use appropriate high-throughput sequencing platforms to perform high-throughput sequencing on a large number of amplicons; (5) Compare the sequenced sequences with the reference database to obtain species annotation information; (6) Perform bioinformatics analysis on the sequencing results.

[0006] In recent years, DNA metabarcoding has made great progress in the study of quantitative dietary composition. Studies have shown that DNA metabarcoding is suitable for determining the food composition of herbivores due to its advantages such as rapidity, simple operation and robustness. However, the analysis of the diet of herbivores based on plant DNA barcodes faces greater difficulties. In the study of the identification of animal food components, people are concerned about whether the metabarcoding technology can make sufficiently accurate estimates of the components consumed by animals. In our previous sheep feeding experiment, we evaluated the accuracy of DNA metabarcoding technology in estimating the dietary composition of grazing livestock. It was found that by amplifying plant ITS2 barcodes from feces and performing Illumina MiSeq sequencing, the types of forage consumed by sheep can be qualitatively identified. And it was preliminarily confirmed that there was a linear relationship between the proportion of forage DNA sequences recovered from sheep feces and the proportion of forage actually consumed by sheep.

[0007] However, due to complex factors such as the types of grass consumed by herbivores and their digestibility, it is inevitable that the measurement results will be biased. The accuracy of the quantitative prediction model previously constructed for the proportion of some forage consumed by sheep still needs to be further supplemented and improved. Summary of the invention

[0008] In one aspect, the present invention provides a method for constructing a livestock feeding composition correction model based on macrobarcodes, comprising: (1) selecting simulated diets derived from different forages as correction samples, extracting mixed sample DNA, performing ITS2 barcode PCR amplification, and performing Illumina MiSeq double-end sequencing of the amplicons to obtain original sequencing data of the correction samples;

[0009] (2) Filter and quality control the raw sequencing data, and cluster the valid sequences into operational taxonomic units (OTUs);

[0010] (3) Compare the OTU representative sequences with the NCBI database Blastn to complete species annotation at the genus or species level;

[0011] (4) Use macrobarcoding technology to measure the forage gene sequence of each mixed sample, obtain the relative proportion of the forage DNA sequence, and analyze and obtain the components of the simulated diet;

[0012] (5) The linear regression method was used to correlate the measured values ​​of each component of the simulated diet and the actual value of the dry matter ratio of each component to establish a correction model for the diet composition.

[0013] Furthermore, extracting DNA from the mixed sample includes a liquid nitrogen grinding step.

[0014] Furthermore, the primers used for PCR amplification of the ITS2 barcode were rD5-ITS2: TCCTCCGCTTATTGATATGC and rb1-ITS2f: CGATACTTGGTGTGAATTGCAG.

[0015] Furthermore, the quality control processing in step (2) includes (i) filtering the bases with a quality value of less than 20 at the end of the read, with 50 bp as a sliding window. If the average quality value of the sequence in the window is less than 20, the back end base is cut off from the window. Reads with a read length less than 50 bp after quality control are filtered; (ii) according to the overlapping relationship between PE reads, the paired reads are spliced ​​into a sequence. The minimum overlap length of the overlapping region between the spliced ​​sequences is 10 bp, and the maximum mismatch ratio allowed in the overlapping region of the spliced ​​sequence is 0.2, and the non-compliant sequences are screened; (iii) the samples are distinguished according to the barcodes and primers at both ends of the sequence, and the sequence direction is adjusted. The number of mismatches allowed for the barcode is 0, and the maximum number of mismatches for the primers is 2. The sequences that cannot be spliced ​​are removed, and finally the barcode label and primer sequences are removed.

[0016] Furthermore, the OTU clustering in step (2) was performed using Usearch version software at 97% similarity, and chimeric sequences were identified and removed using UCHIME software during the clustering process.

[0017] Furthermore, a regression correction step is also included.

[0018] Furthermore, the Kulczyński similarity coefficient (KSI, %) was used in regression correction to evaluate the accuracy, and its calculation formula is as follows:

[0019]

[0020] Among them, c i is the smaller value between the actual value and the estimated value of the i-th forage in the simulated diet; (a i +b i) is the sum of the ratios of the true value to the estimated value of each forage component.

[0021] On the other hand, the present application provides the use of the above method in determining the feed composition of livestock.

[0022] Furthermore, the livestock is sheep.

[0023] The forage grass in the present application can be selected from various known or unknown species according to specific needs and conditions, including but not limited to 9 genera including Puccinellia, Cynodon, Poa, Lolium, Bromus, Leymus sp., Medicago, Trifolium, Cichorium, and specific species including Puccinellia distans, Cynodon dactylon, Poa pratensis, Lolium perenne, Bromus erectus, Medicago sativa, Trifolium pratense, Cichorium intybus, Trifolium rubens, Trifolium squamosum, Bromushordeaceus, etc.

[0024] The livestock in this application include various herbivorous or omnivorous livestock, including but not limited to sheep such as goats, sheep or meat sheep, wool sheep, and meat and wool sheep; cattle such as dairy cows and beef cattle; pigs; horses; donkeys, etc.

[0025] The molecular biology operation methods in this application, such as DNA extraction and PCR methods, can be selected and implemented by those skilled in the art according to conventional technical means in the art (including but not limited to Molecular Cloning, Sambrook and other reference books), and various finished kits can be selected as needed.

[0026] The beneficial effects of the present invention include: the present invention establishes a regression correction model for simulated dietary forage components, including an overall regression equation and a single regression equation. By comparing with the actual proportions of the simulated dietary components, the accuracy of the estimated values ​​and the corrected values ​​is analyzed using the similarity coefficient KSI, and it is found that after correction with the overall regression equation and the single regression equation, the accuracy of the measurement results of the edible components is significantly improved. The constructed model is used to correct the edible components of house-fed sheep, and the accuracy of the measurement results is significantly improved after correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is the relative abundance (genus level) of forage in the simulated diet;

[0028] Figure 2 To simulate the relative abundance of forage in the diet (species level);

[0029] Figure 3 To simulate the regression relationship between the estimated and true values ​​of forage components in the diet;

[0030] Figure 4 To simulate the linear regression model of each forage species in the diet;

[0031] Figure 5 The similarity coefficient (KSI) between the estimated values ​​of forage components based on the metabarcoding technology and the actual values ​​in the simulated diet treatments. The estimated values ​​of forage components include uncorrected (gray squares), overall regression equation (CE-1, red circles) and individual regression equations (CE-2, blue triangles). D: simulated diet; RC: regression correction; D×R: interaction between simulated diet and regression correction method. Asterisks*** indicate significant differences at the P<0.001 level;

[0032] Figure 6 The coefficient of similarity (KSI) between the estimated and measured values ​​of the dietary components based on DNA barcoding technology in dietary treatments 1, 2, 3, and 4. The estimated values ​​of the dietary components detected in feces include uncorrected (gray circles), correction equation 1 (black squares), and correction equation 2 (black triangles). D: diet; RC: regression correction; D×R: interaction between diet and regression correction method. Asterisks*** indicate significant differences at the P<0.001 level;

[0033] Figure 7 Comparison of the similarity coefficients (KSI) between the estimated values ​​of feed composition based on the metabarcoding and saturated alkanes methods and the true values ​​in diet treatments 1, 2, 3, and 4. The similarity coefficients were calculated based on the correction methods including uncorrected (gray circles), correction equation 1 (black squares), correction equation 2 (blue triangles), individual recovery (black circles), treatment recovery (red circles), and overall recovery (black triangles). D: diet; CM: correction method; D×CM: interaction between diet and correction method. Asterisks*** indicate significant differences at the P<0.001 level. DETAILED DESCRIPTION

[0034] Example 1 Determination of forage components in simulated diets

[0035] The nine types of forage grasses collected were fully ground with liquid nitrogen and mixed thoroughly according to the set ratio (dry matter ratio) as simulated diets. The ratios are shown in Table 1 below. There were 10 treatments, each with 4 replicates, for a total of 40 mixed forage samples. About 0.5 g of each mixed forage sample was placed in a centrifuge tube and stored in a liquid nitrogen tank.

[0036] Table 1 Composition of simulated diet

[0037]

[0038] The simulated diet samples were ground and mixed with liquid nitrogen and the total DNA of the sample plants was extracted using a plant genomic DNA extraction kit (DNAsecurePlant Kit). The above DNA was used as a template to PCR amplify the plant ITS2 sequence using primers rD5-ITS2 and rb1-ITS2f (Table 2) reported in the literature, and an 8bp tag sequence was added to each sample to distinguish the sample source of each data during sequencing analysis.

[0039] Table 2 Illumina MiSeq sequencing primers

[0040]

[0041] PCR amplification reaction was performed on ABI The PCR amplification conditions of ITS2 gene were as follows: 94℃ pre-denaturation for 5 minutes; 94℃ denaturation for 30 seconds, 59℃ annealing for 1 minute, 72℃ extension for 1 minute; reaction for 45 cycles; 72℃ extension for 10 minutes. A blank control should be set for each PCR reaction to exclude sample amplification contamination. The amplification of each sample was repeated three times, and the reaction system was 20μL, including 4μL of 5×FastPfu buffer, 2μL of 2.5mM dNTPs, 0.8μL of each primer (5μM), 0.4μL of FastPfu Polymerase, 0.2μL of Bovine Serum Albumin (BSA) and 10ng of DNA template. All PCR amplification products were observed by electrophoresis on 2.0% agarose gel, and the band size was compared using DL2000 marker. Positive PCR products were selected for the next sequencing analysis.

[0042] The PCR products were recovered from 2% agarose gel using the AxyPrep DNA gel recovery kit (Axygen Bioscience, Union City, CA, US) and detected and quantified using the QuantiFluorTM-ST blue fluorescence quantitative system (Promega, US). The purified PCR amplicons were then mixed in equal proportions. Double-end (2×300 bp) sequencing was then performed according to the standard process of the Illumina MiSeq sequencing platform.

[0043] The original fastq files were strictly filtered and quality-controlled using QIIME version 1.17 software. The data filtering methods and parameters were as follows: (i) Filter bases with a quality value of less than 20 at the end of the read, with a sliding window of 50 bp. If the average quality value of the sequence in the window is less than 20, the back-end bases are truncated from the window. Filter sequences (reads) with a read length less than 50 bp after quality control; (ii) According to the overlap relationship between PE reads, the paired reads were spliced ​​into a sequence. The minimum overlap length between the spliced ​​sequences was 10 bp. The maximum mismatch ratio allowed in the overlap region of the spliced ​​sequence was 0.2, and the non-compliant sequences were screened; (iii) Distinguish samples according to the barcode and primers at both ends of the sequence, and adjust the sequence direction. The number of mismatches allowed for the barcode was 0, and the maximum number of primer mismatches was 2. Remove the sequences that cannot be spliced, and finally remove the barcode label and primer sequences.

[0044] After filtering and quality control, the original sequence file forms a new sequence file, and each sequence has a specific label. The sequences were clustered into operational taxonomic units (OTUs) according to 97% similarity using Usearch version 7.1 software. During the clustering process, the chimeric sequences were identified and removed using UCHIME software to obtain the representative sequence of OTU. All optimized sequences were matched to the representative sequence of OTU, and sequences with a similarity of more than 97% to the representative sequence of OTU were selected to generate the OTU table.

[0045] During the taxonomic comparison, the GenBank "nucleotide collection (nr / nt)" database was used, and the online BLASTN tool was used to annotate each OTU according to the principle of highest sequence homology. If two or more representative OTU sequences had the same score or were not aligned to the species level, the sequence was annotated at a higher level (genus or family level) for species taxonomy.

[0046] The irrelevant sequences (fungi) in the OTU table were removed, and then the simulated diet composition was analyzed. OTU sequences with a number greater than or equal to 2 were identified as valid sequences. Sequences with a number less than 2 may be caused by ITS2 sequencing errors and are not considered in subsequent analysis. After logarithmic transformation of the DNA sequence number, the relative proportion of each forage DNA sequence in the simulated diet sample was calculated, and then the components of the simulated diet were analyzed.

[0047] After sequencing the ITS2 barcodes in 40 mixed forage samples (simulated diet), a total of 1,473,296 high-quality sequences were obtained. After clustering and removing fungi, 40 OTUs were obtained. The valid sequences in all samples were classified and analyzed, and the forage species and relative abundance corresponding to the taxonomic unit were counted at the genus and species levels ( Figure 1 and Figure 2 ). The OTU representative sequences were annotated, and the results showed that there were 9 genera in the samples, including Puccinellia, Cynodon, Poa, Lolium, Bromus, Leymus sp., Medicago, Trifolium, and Cichorium. The three genera with the highest abundance were Trifolium, Cichorium, and Medicago, accounting for 62.5% of the total OTUs. Among them, 36 OTUs were identified to species level, including 11 plant species, including Puccinellia distans, Cynodon dactylon, Poa pratensis, Lolium perenne, Bromus erectus, Medicago sativa, Trifolium pratense, Cichorium intybus, Trifolium rubens, Trifolium squamosum and Bromus hordeaceus.

[0048] Example 2 Regression model construction and accuracy test of simulated diet components

[0049] Linear regression analysis was used to determine the quantitative relationship between the relative proportions of DNA sequences obtained by sequencing and the proportion of forage in the simulated diet, and the resulting regression equation was used to correct the direct estimation results.

[0050] The direct estimation results of the food intake components obtained by high-throughput sequencing were subjected to regression correction, and the accuracy of the estimated and corrected results was evaluated using the Kulczyński similarity coefficient (KSI, %). The calculation formula is as follows:

[0051]

[0052] Among them, c i is the smaller value between the actual value and the estimated value of the i-th forage in the simulated diet; (a i +b i ) is the sum of the ratios of the true value to the estimated value of each forage component.

[0053] Illumina MiSeq sequencing was used to obtain the number of sequences for each OTU. After logarithmic transformation, the relative proportions of each forage in different mixed forage samples were calculated and then compared with the actual proportions in the mixed forage samples. Then, a regression analysis was performed on the relative proportions of the DNA sequences of the nine forage species and their actual proportions in the simulated diet. The results showed that there was a very significant linear regression relationship between the two ( Figure 3 ). SPSS software was used to fit the parameters and the overall regression prediction equation of the forage components in the simulated diet was obtained as y=0.425x+0.064.

[0054] A total of 10 simulated diets with different proportions were set up, including 9 common grass species. Individual linear regression equations were established for each of the 9 grass species and variance analysis was performed, such as Figure 4 These equations are: Puccinellia tenuifolia (PD), y = 0.946x + 0.098; Bermuda grass (CD), y = 0.999x + 0.088; Poa annua (PP), y = 0.617x + 0.044; Lolium perenne (LP), y = 0.377x + 0.022; Bromus esculenta (BE), y = 0.437x + 0.038; Leymus chinensis (LC), y = 0.650x + 0.059; Alfalfa (MS), y = 1.727x - 0.189; Red clover (TP), y = 1.018x - 0.065; Chicory (CI), y = 0.055x + 0.095.

[0055] The relative proportions of various forages in the diet were estimated by DNA macrobarcoding technology, and the estimated results were regressed using two models: the overall regression equation (CE-1) and the single regression equation (CE-2). Table 1 shows the actual measured values ​​(Oberserved), estimated values ​​(Estimated), and regression-corrected values ​​(CE-1 and CE-2) of the nine forages in the simulated diet. The similarity coefficient KSI was used to evaluate the accuracy of the estimated results and the results after correction by the two methods. Figure 5The following are the variance analysis results of the similarity coefficient KSI of the three results. It can be seen that the dietary treatment (D), correction method (RC) and the interaction between the two (D×RC) have extremely significant effects on KSI (P<0.001).

[0056] Obviously, after correction by single regression equation and overall regression equation, the accuracy of the estimated results of dietary components was significantly improved. The KSI values ​​of the direct estimation results (uncorrected) were 66.02% to 76.18%, with an average of 70.15%; the KSI values ​​corrected by the overall regression equation (CE-1) were 73.84% to 86.37%, with an average of 80.42%; the KSI values ​​corrected by the single regression equation (CE-2) were 85.84% to 93.29%, with an average of 90.58%.

[0057]

[0058]

[0059] Example 3 Estimation and calibration of diet composition for confined sheep

[0060] The ITS2 barcode sequencing method of the present invention was used to estimate the dietary composition of sheep. The KSI was also used to test the accuracy of the method ( Figure 6 ). The results of variance analysis showed that dietary treatment, regression equation correction method and the interaction between the two had a very significant effect on KSI (P<0.001). After correction by two regression equations, KSI increased significantly (P<0.001). Comparing the three estimation results, it can be found that after correction by the second correction equation (CE-2), KSI was the highest (88.41%), followed by the second correction equation CE-1 (70.24%), and the uncorrected KSI value was the lowest (57.97%).

[0061] From the perspective of dietary composition, the similarity between the estimated and true values ​​of the dietary composition directly estimated using DNA macrobarcoding technology was low (46.98% to 67.99%), and the KSI decreased significantly with the increase of dietary composition. After correction by CE-1 (66.89% to 72.03%) and CE-2 (84.5% to 92.0%), the KSI was relatively stable with the increase of dietary composition, and did not show a sudden downward trend.

[0062] Example 4 Comparison of application effects with saturated alkane technology

[0063] In this study, the diet composition of sheep was estimated using metabarcoding and saturated alkane technology, and the accuracy of the two technologies was evaluated using the similarity coefficient KSI. Figure 7The similarity between the estimated values ​​and the true values ​​of the dietary components of sheep under four dietary treatments using six methods is presented. The results show that the accuracy of the six estimated results is in the following order: individual recovery (I) > treatment recovery (T) > calibration equation 2 (CE-2) > overall recovery (G) > calibration equation 1 (CE-1) > direct estimate (estimated). Table 2 lists the multiple comparison results of the six calibration methods. It can be seen that the similarity coefficient KSI does not differ significantly between calibration method 2 and treatment recovery.

[0064] The differences in the recoveries of indicators in feces increased from individual sheep recoveries, treatment recoveries to overall recoveries, and the accuracy of the estimated results of the dietary composition was improved after the recovery correction. However, since the dietary composition significantly affected the recovery of long-chain alcohols in feces, the overall recovery of long-chain alcohols caused by dietary composition deviated greatly from the average recovery of the treatment. Therefore, the overall recovery was only accurately estimated for one of the dietary components. The estimate corrected for individual sheep recoveries was the most accurate, but it was difficult to operate in the field. Compared with the method of single sheep recovery correction, although the accuracy of the estimate corrected for the same dietary treatment recovery was reduced, the impact was small. Therefore, it is recommended to use the same dietary treatment recovery correction method.

[0065] Similarly, the similarity coefficient KSI was used to evaluate the estimation results of sheep diet composition using DNA barcoding technology. First, we used two correction equations to correct the estimation results. After the regression equation correction, it was found that the KSI value increased significantly compared with that before correction, that is, the accuracy of the estimation results was significantly improved ( Figure 7 ). Moreover, the results are different from those of the saturated alkane method, that is, as the intake of dietary components increases, the change range of KSI is relatively small, and it is less affected by the intake of dietary components. Therefore, it can be predicted that for the estimation of complex components, the macrobarcoding technology is more accurate than the saturated alkane method.

[0066] In the research results of saturated alkane technology, from the perspective of practical application, we recommend using the treatment recovery rate to correct the alkane recovery rate. Figure 7 The variance analysis results in Table 2 show that after the results estimated by the macrobarcode technology are corrected by the second method, there is no significant difference in the accuracy of the two, which also confirms the application effect and potential of the present invention in estimating the feeding components of sheep.

[0067] Table 4 Similarity coefficients between the estimated values ​​and the true values ​​of the six correction methods

[0068]

[0069] This embodiment is only a preferred specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for constructing a macrobarcoding-based livestock feed composition correction model: It is characterized in that It includes: (1) Simulated diets derived from different forages were selected as calibration samples, and DNA of mixed samples was extracted, ITS2 barcode PCR amplification was performed, and the amplicons were sequenced by Illumina MiSeq double-end sequencing to obtain the original sequencing data of the calibration samples; (2) Filter and quality control the raw sequencing data, and cluster the valid sequences into operational taxonomic units (OTUs); (3) Compare the OTU representative sequences with the NCBI database Blastn to complete species annotation at the genus or species level; (4) Use macrobarcoding technology to measure the forage gene sequence of each mixed sample, obtain the relative proportion of the forage DNA sequence, and analyze and obtain the components of the simulated diet; (5) The linear regression method was used to correlate the measured values ​​of each component of the simulated diet and the true value of the dry matter ratio of each component to establish a correction model for the diet composition. The KSI similarity coefficient was used to evaluate the accuracy, and its calculation formula is as follows: Where ci is the smaller of the true and estimated values ​​of the ith forage in the simulated diet; (ai+bi) is the sum of the ratios of the true and estimated values ​​of each forage component. The primers used for PCR amplification of the ITS2 barcode in step (1) are rD5-ITS2: TCCTCCGCTTATTGATATGC and rb1-ITS2f: CGATACTTGGTGTGAATTGCAG; The quality control processing in step (2) includes (i) filtering bases with a quality value of less than 20 at the end of the read, using 50 bp as a sliding window. If the average quality value of the sequence in the window is less than 20, the back-end base is truncated from the window; filtering reads with a read length less than 50 bp after quality control; (ii) splicing paired reads into a sequence based on the overlapping relationship between PEreads. The minimum overlap length of the overlapping region between the spliced ​​sequences is 10 bp, and the maximum mismatch ratio allowed in the overlapping region of the spliced ​​sequence is 0.2, and non-compliant sequences are screened; (iii) distinguishing samples based on the barcodes and primers at both ends of the sequence, and adjusting the sequence direction. The barcode mismatch number allowed is 0, and the maximum primer mismatch number is 2. Sequences that cannot be spliced ​​are removed, and finally the barcode label and primer sequence are removed; OTU clustering in step (2) uses Usearch version 7.1 software, according to 97% similarity, and UCHIME software is used to identify and remove chimeric sequences during the clustering process.

2. The construction method according to claim 1, wherein extracting the mixed sample DNA comprises a liquid nitrogen grinding step.

3. Use of the method according to claim 1 or 2 in determining the feed composition of livestock.

4. The use according to claim 3, wherein the livestock is sheep.