A liquid chip detection method for the identification of Leymus chinensis varieties

Abstract

The present invention belongs to the fields of biotechnology and agricultural science, and particularly relates to a liquid-phase chip detection method for the identification of Leymus chinensis varieties. The method of the present invention uses a 50K liquid-phase chip of the whole genome of Leymus chinensis for the identification of Leymus chinensis varieties. This method combines high-throughput liquid-phase chip technology with machine learning algorithms, and through high-throughput SNP detection, realizes the accurate detection of SNP sites in the whole genome, innovatively constructs an efficient and stable classification model, and significantly improves the accuracy and efficiency of variety identification. By screening specific SNP sites, combining population genetic analysis with the random forest algorithm, the reliability of the model is enhanced, and problems such as low efficiency of traditional methods are solved.

Description

Technical Field

[0001] The present invention belongs to the fields of biotechnology and agricultural science, and particularly relates to a liquid-phase chip detection method for the identification of Leymus chinensis varieties. Background Art

[0002] Leymus chinensis ( Leymus chinensis (Trin.) Tzvel.) belongs to the perennial forage grass of the genus Leymus in the family Poaceae, and is one of the important constructive species in the meadow steppe and arid steppe in the eastern part of the Eurasian steppe region. The Songnen Plain in the northeast of China and the eastern part of Inner Mongolia are also the main distribution centers of Leymus chinensis. At the same time, it is widely distributed in Hebei, Shanxi, Henan, Shaanxi, Ningxia, Gansu, Qinghai, Xinjiang and other regions; in addition, it is also sporadically distributed in Russia, Mongolia, North Korea, Japan and other places. As an important native forage grass in northern China, Leymus chinensis not only has high nutritional value, stable yield and good palatability, and is liked by various livestock, but also has excellent stress resistance and can adapt to extreme climates and diverse soil environments. Moreover, the well-developed horizontal rhizomes of Leymus chinensis endow it with strong penetration and invasion ability, and by forming a strong root network, it has a significant role in soil and water conservation. Therefore, Leymus chinensis plays an irreplaceable role in the ecological construction of northern grasslands, vegetation restoration, and the development of animal husbandry.

[0003] In recent years, with the degradation of grassland ecology and the increasing demand for high-quality forage grass in animal husbandry, the breeding of new Leymus chinensis varieties and the protection of resources have become research hotspots. However, due to its genetic diversity and morphological complexity, traditional classification and identification methods are difficult to meet the requirements of rapidity and accuracy. At present, there is still a blank in the efficient molecular markers and genotype detection technologies for Leymus chinensis, which restricts the progress of genetic research and molecular breeding. Therefore, the development of efficient molecular tools suitable for the identification of Leymus chinensis varieties is of great significance for promoting its research and application.

[0004] With the rapid development of molecular biology and genomics technologies, traditional phenotypic identification methods are gradually being replaced by identification means at the molecular level. Molecular marker technologies (such as SSR, RAPD, AFLP, etc.) and high-throughput detection platforms have been widely used in forage grass breeding and resource evaluation. Among them, the molecular detection technology based on SNP (single nucleotide polymorphism) has become an important tool for plant genetic research and molecular breeding due to its high efficiency, sensitivity and high-throughput characteristics. As a highly stable molecular marker widely existing in the genome, SNP is very suitable for large-scale and high-throughput genotype analysis. And the SNP liquid-phase chip is a genotyping technology platform based on high-throughput detection, which can detect thousands to tens of thousands of SNP loci at one time, and has high sensitivity, low cost and strong flexibility. At present, the SNP liquid-phase chip has been successfully applied to variety identification and molecular breeding of crops including rice, corn, wheat and various forage grasses, providing strong technical support for agricultural and forage grass improvement.

[0005] However, although SNP liquid chip technology has achieved remarkable results in crop and forage research, the identification of Leymus chinensis varieties still mainly relies on traditional morphological indicators. These traditional methods require long-term observation, are susceptible to environmental factors and subjective human judgment, and have problems such as long identification cycles, low accuracy, and strong subjectivity of results. In addition, as an allopolyploid plant, Leymus chinensis has a complex genome structure, a relatively large genome size, and a high proportion of repetitive sequences. Compared with other major crops and forages, the genomic research of Leymus chinensis lags behind. Currently, advanced technologies such as SNP liquid chips have not been introduced into the molecular research of Leymus chinensis, and the lack of high-throughput molecular detection and genotype analysis tools limits the efficiency and accuracy of Leymus chinensis variety identification. Therefore, there is an expectation in this field to develop a method for molecular identification of Leymus chinensis varieties based on liquid chip technology, which can achieve rapid and accurate identification of Leymus chinensis varieties and has positive significance for the protection and utilization of Leymus chinensis germplasm resources. Summary of the Invention

[0006] For this reason, the purpose of the present invention is to provide a liquid chip detection method for Leymus chinensis variety identification. By screening specific SNP sites, combining population genetic analysis and random forest algorithm, the reliability of the model is enhanced to solve problems such as low efficiency of traditional methods.

[0007] To solve the above technical problems, a liquid chip detection method for Leymus chinensis variety identification according to the present invention includes the following steps:

[0008] (1) Perform SNP detection on different varieties of Leymus chinensis using a 50K liquid chip of the Leymus chinensis whole genome. Among them, the sites of the liquid chip include multiple SNP sites located on the Leymus chinensis reference genome Leymus_chinensis_Lc6-5 or its updated version, covering gene regions and their related non-gene regions;

[0009] The specific information of the core SNP sites of the liquid chip is shown in Table 1 of the specification; among them, the position and variation information of the SNP sites are represented in the form of chromosome / physical position / reference genotype / variant allele genotype, and the physical position of the SNP molecular marker combination is located and analyzed based on the Leymus chinensis reference genome;

[0010] (2) Compare the SNP sequencing results with the Leymus chinensis reference genome, perform SNP analysis and annotation, and establish a SNP information library for different varieties.

[0011] Furthermore, for the liquid chip detection method for Leymus chinensis variety identification, SNPs identified by sequencing are used, and the distance matrix is calculated using the TreeBeST software. Based on this, a phylogenetic tree of 11 varieties is constructed by the neighbor-joining method. The GCTA software is used for principal component analysis (PCA), and based on the SNP differences of individual genomes, individuals are clustered into different subgroups according to different trait characteristics. The software admixstructure is used to analyze the population genetic structure of the samples.

[0012] Furthermore, for the liquid chip detection method for Leymus chinensis variety identification, the SNP data set obtained by sequencing is encoded, and then the principal SNP loci are extracted using PCA and singular value decomposition. After PCA preselection and feature filtering, the number of SNP loci is still very large. To evaluate the importance of the selected SNP loci, the samples are assigned to the corresponding varieties, and random forest is used for variety classification.

[0013] Furthermore, for the liquid chip detection method for Leymus chinensis variety identification, the optimal number of decision trees shows a monotonic increasing trend in the accuracy of the random forest model, but after reaching a certain value, the effect tends to be stable or even fluctuate, and at the same time, the computational cost increases significantly. Therefore, the optimal number of trees is determined after balancing accuracy and computational cost. Random forest uses the bootstrap sampling method to generate decision trees, and evaluates the classification performance through the out-of-bag (OOB) data that is not drawn, obtaining an unbiased OOB error rate or accuracy rate. In addition, to further verify the stability and reliability of the model, k-fold cross-validation is also carried out.

[0014] Furthermore, for the liquid chip detection method for Leymus chinensis variety identification, first, the trained model is used to score the OOB data, and the AUC or other evaluation indicators are calculated. Then, for each feature in the OOB data, the following operations are performed in sequence: randomly shuffle the values of this feature to observe the change in model performance; re-score and calculate the evaluation indicators; calculate the change rate of the indicators. Through the above steps, the change rate of each feature is obtained, and the features are sorted according to the change rate, thereby quantifying the feature importance (i.e., locus importance). According to the importance ranking, the feature importance is accumulated, and the loci with the accumulated value exceeding 90% are selected, and the random forest model is re-trained, and finally 615 SNP loci are screened out. Finally, the accuracy of the model is tested using this model, and cross-validation is carried out.

[0015] Furthermore, for the liquid chip detection method for Leymus chinensis variety identification, to further verify the model performance, each population is randomly divided into a training set and a test set in a ratio of 7:3, and iterated 100 times to evaluate the prediction accuracy. In addition, the screened SNP locus information is encoded as a two-dimensional code, and users can obtain the molecular identity information of the corresponding variety by scanning the two-dimensional code.

[0016] Further, the liquid-phase chip detection method for Leymus chinensis variety identification further includes the step of identifying the variety of the to-be-detected Leymus chinensis, including the step of obtaining the genotype data of the to-be-detected Leymus chinensis individual based on the 50K liquid-phase chip of the whole Leymus chinensis genome, and the step of comparing and identifying the Leymus chinensis variety based on the constructed reference sample library.

[0017] Further, for the liquid-phase chip detection method for Leymus chinensis variety identification, the 50K liquid-phase chip of the whole Leymus chinensis genome includes a separately packaged 50K probe mixture and reagents suitable for liquid-phase hybridization capture; wherein,

[0018] The probe mixture contains 50K probes with high sequence specificity, which can identify and detect the core SNP molecular marker sites as shown in Table 1 of the specification, and the range of SNP sites that the probes can identify and detect includes but is not limited to the sites listed in Table 1.

[0019] As an exemplary embodiment, the 50K liquid-phase chip of the whole Leymus chinensis genome includes a separately packaged 50K probe mixture and a Hyb&Wash kit suitable for liquid-phase hybridization capture; wherein,

[0020] The probe mixture contains 50K probes with high sequence specificity, which can effectively identify and detect the core SNP molecular marker sites as shown in Table 1 of the specification.

[0021] Specifically, for the liquid-phase chip detection method for Leymus chinensis variety identification, the 50K liquid-phase targeted capture probes are designed based on the positive strand of the Leymus chinensis reference genome;

[0022] Preferably, the designed length range of the probes is 55 - 120 nt, and the target average length is 100 nt.

[0023] Preferably, the 50K liquid-phase targeted capture probes are designed based on the principle of thermodynamic stability to ensure 99% capture in the conventional region;

[0024] Preferably, for the 50K liquid-phase targeted capture probes, for the difficult and complex regions, by adjusting the probe positions and adopting the scheme of placing multilayer "overlapping" probes, the effective coverage of the difficult regions is improved, and more preferably, it is a 3X overlapping structure.

[0025] The method for Leymus chinensis variety identification described in the present invention utilizes a 50K liquid-phase chip for the whole genome of Leymus chinensis. This method combines high-throughput liquid-phase chip technology with machine learning algorithms. Through high-throughput SNP detection, it realizes accurate detection of SNP sites across the whole genome, innovatively constructs an efficient and stable classification model, and significantly improves the accuracy and efficiency of variety identification. By screening specific SNP sites, combining population genetic analysis and random forest algorithms, it enhances the reliability of the model and solves problems such as low efficiency of traditional methods. Using specific SNPs to generate two-dimensional codes, it realizes variety traceability and promotion, providing scientific support for the protection of forage germplasm resources and the development of agriculture and animal husbandry.

[0026] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention performed library construction and sequencing on a total of 228 samples of 11 Leymus chinensis varieties (including 9 varieties, 1 strain, and 1 germplasm), compared them with the Leymus chinensis reference genome, and conducted SNP detection and annotation, realizing rapid and accurate identification of Leymus chinensis varieties. It not only overcomes the limitations of traditional methods but also provides important technical support for grassland ecological restoration and the development of the forage industry. At the same time, it promotes the application and promotion of high-quality new varieties, laying a foundation for the protection and utilization of Leymus chinensis germplasm resources.

[0027] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on a 50K liquid-phase chip for the whole genome of Leymus chinensis for detection. The SNP molecular marker combination of Leymus chinensis was determined by whole-genome resequencing. A total of 51,696 SNP molecular marker sites were screened based on the Leymus chinensis reference genome Leymus_chinensis_Lc6-5. The 50K liquid-phase chip for the whole genome of Leymus chinensis was developed based on targeted capture sequencing technology. By combining high-density SNP site probes, it realizes high-throughput and high-precision genotyping of the Leymus chinensis genome, and can comprehensively evaluate the genetic characteristics of Leymus chinensis.

[0028] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on a 50K liquid-phase chip for the whole genome of Leymus chinensis for detection. In addition to using the Leymus chinensis reference genome, de novo transcriptome homology alignment analysis is simultaneously adopted for probe design to optimize SNP site selection and genome coverage, ensuring the uniform distribution and good polymorphism of SNP sites in the Leymus chinensis genome. The probe sites cover key functional gene regions of Leymus chinensis, including genes for important traits such as growth, development, and resistance, and can provide important bases for molecular marker identification of agronomic traits, quality traits, and stress resistance traits of Leymus chinensis.

[0029] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on the Leymus chinensis whole-genome 50K liquid-phase chip for detection. It is designed based on the principle of thermodynamic stability to ensure 99% capture of the conventional region. For the difficult and complex regions, by adjusting the probe positions and adopting the multi-layer "overlapping" probe placement scheme, the effective coverage of the difficult regions is improved. At the same time, it ensures the uniform distribution of SNP sites in the genome and the coverage of key functional regions, improving the coverage of functional regions and supporting the research on agronomic traits and stress resistance traits.

[0030] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on the Leymus chinensis whole-genome 50K liquid-phase chip for detection, which can provide high-quality genotyping data. Compared with other genotyping technologies (such as GBS), the data consistency is higher, and it has stronger data comparison and accumulation capabilities. The high throughput and data consistency of this chip provide a strong guarantee for long-term data storage and comprehensive utilization, greatly improving the long-term availability of data and the simplicity of analysis.

[0031] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on the Leymus chinensis whole-genome 50K liquid-phase chip for detection. The liquid-phase chip technology has higher flexibility compared with traditional solid-phase chips. It can adjust the SNP density of the chip according to needs, provide multiple versions such as 50K, 40K, 30K, etc., and can capture the sequence information of 100 bp upstream and downstream of the target SNP site, providing more data for functional gene research and adapting to different scales of research and application requirements. The liquid-phase chip can complete high-throughput detection, has high detection flexibility, can effectively shorten the sample detection cycle, and improve scientific research efficiency.

[0032] The liquid-phase chip detection method for Leymus chinensis variety identification described in the present invention is based on the Leymus chinensis whole-genome 50K liquid-phase chip for detection. Based on the targeted capture sequencing technology, the tolerance of the flanking sequence is fully considered in probe design. When the variation of the flanking sequence does not exceed 10%, the target sequence can still be stably captured. This technical characteristic ensures the high reliability of the data. At the same time, in addition to obtaining the information of the target SNP site, this chip can also obtain the sequence information of 100 bp upstream and downstream each, providing more supporting data for the functional gene research of Leymus chinensis.

[0033] The Leymus chinensis whole-genome 50K liquid-phase chip described in the present invention realizes the completely domestic synthesis of probes, sample detection and reagent use, avoiding the high costs brought by imported equipment and reagents. By reducing material costs and reducing technical dependence, it realizes the autonomy and controllability of production and application, reduces the risk of core data leakage, and avoids the inconvenience brought by external trade frictions, meeting the domestic demand for the autonomy and controllability of agricultural science and technology. Description of the Drawings

[0034] In order to make the content of the present invention easier to be clearly understood, the following further details the present invention according to specific embodiments of the present invention in combination with the accompanying drawings, wherein,

[0035] Figure 1 It is the flow chart of SNP locus screening in Example 1 of the present invention;

[0036] Figure 2 It is the distribution map of 51,696 SNP locus probes on chromosomes in Example 3 of the present invention;

[0037] Figure 3 It is the statistical chart of the distribution of SNP markers in the gene structure in Example 3 of the present invention;

[0038] Figure 4 It is the radar chart of the expected performance and actual test performance of the Leymus chinensis 50K liquid chip in Example 4 of the present invention;

[0039] Figure 5 It is the flow chart of library construction and sequencing of the Leymus chinensis liquid chip in Example 5 of the present invention;

[0040] Figure 6 It is the annotation result map of the SNP loci detected in Example 6 of the present invention;

[0041] Figure 7 It is the result map of the population evolutionary tree of Leymus chinensis varieties in Example 6 of the present invention;

[0042] Figure 8 It is the result map of the population principal component analysis of Leymus chinensis varieties in Example 6 of the present invention;

[0043] Figure 9 It is the population genetic structure map of Leymus chinensis varieties in Example 6 of the present invention;

[0044] Figure 10 It is the model prediction accuracy map in Example 7 of the present invention, taking breed_11 as an example;

[0045] Figure 11 It is the SNP fingerprint information of Leymus chinensis varieties in Example 7 of the present invention, taking breed_11 as an example. Specific embodiments

[0046] In the following embodiments of the present invention, the liquid chip detection method for Leymus chinensis variety identification, the Leymus chinensis whole-genome 50K liquid chip is developed based on the target capture technology, and its working principle is to design probes for the target region sequence, target and capture the target fragment, and then use the sequencing platform for high-throughput sequencing to target detect the target genes and loci, quickly detect a large number of samples, so as to achieve the purpose of target region gene detection and genotyping.

[0047] In the following embodiments of the present invention, the 50K liquid-phase chip of the whole genome of Leymus chinensis contains a set of 50K SNP probe sites of Leymus chinensis.

[0048] In the following embodiments of the present invention, the 50K liquid-phase chip of the whole genome of Leymus chinensis includes a separately packaged 50K probe mixture and a Hyb&Wash kit suitable for liquid-phase hybridization capture.

[0049] In some embodiment solutions, the liquid-phase chip of the present invention uses the principle of nucleic acid hybridization for targeted capture. When two nucleic acid strands are complementary, a stable double-stranded structure will form between them. Through the denaturation and renaturation processes, first denature the target region library to remove repetitive sequences and linker sequences, and then add probes complementary to the bases of the target region for renaturation; the non-specific binding is removed by washing, and finally a specific probe-library conjugate is obtained, thus completing the capture of the target region.

[0050] In some embodiment solutions, the preparation of the liquid-phase targeted capture probe of the liquid-phase chip of the present invention includes template synthesis and RNA preparation. The probe template is composed of single-stranded DNA (oligo) synthesized organically. After quality control (QC) and pooling, the qualified template is used to prepare RNA analogue probes. In the synthesis of RNA probes, biotin-labeled nucleotide analogues (such as LNA, PNA, etc.) are used to improve the binding stability between the probe and the library. The biotin-labeled probe can bind to streptavidin-coated magnetic beads to further improve the capture efficiency.

[0051] In some embodiment solutions, for the 50K liquid-phase chip of the whole genome of Leymus chinensis, the 50K liquid-phase targeted capture probe is designed based on the positive strand of the Leymus chinensis reference genome;

[0052] Preferably, the designed length range of the probe is 55 - 120 nt, and its target average length is 100 nt;

[0053] Specifically, the probe design of this liquid-phase chip is based on the principle of thermodynamic stability, including factors such as melting temperature (Tm), GC content, and sequence specificity, and comprehensively considers the genomic complexity of the target species, the position of the target site, and the GC content around it to ensure 99% capture of the conventional region. For complex regions with too high or too low GC content, by adjusting the probe position, a multi-layer "tiling" type probe design is adopted to improve the coverage of difficult regions. This solution improves the capture efficiency in extreme GC regions by increasing the complementarity and stability between probes. The probe design is based on the positive strand of the latest version of the Leymus chinensis reference genome. The designed length range of the probe is 55 - 120 nt, and its target average length is 100 nt, which is optimized to ensure efficient SNP capture.

[0054] In the following embodiments of the present invention, the liquid-phase chip detection method for Leymus chinensis variety identification utilizes a 50K liquid-phase chip of the Leymus chinensis whole genome, combines high-throughput liquid-phase chip technology with machine learning algorithms, and through high-throughput SNP detection, realizes accurate detection of SNP sites across the whole genome, innovatively constructs an efficient and stable classification model, and significantly improves the accuracy and efficiency of variety identification. By screening specific SNP sites, combining population genetic analysis with the random forest algorithm, the reliability of the model is enhanced, and problems such as low efficiency of traditional methods are solved. Using specific SNPs to generate two-dimensional codes realizes variety traceability and promotion.

[0055] Example 1

[0056] This example provides a 50K liquid-phase chip of the Leymus chinensis whole genome, and the specific screening process is as Figure 1 shown, and the specific screening process is as follows.

[0057] (I) Preliminary filtering and screening of SNPs

[0058] In this example, 100 Leymus chinensis samples were re-sequenced, and SNP variant detection was performed on these sequencing data. The reference genome version used was Leymus_chinensis_Lc6-5. The preliminary SNP filtering and screening process is as follows:

[0059] (1) Depth filtering: On the basis that the average sequencing depth of the samples reaches 10X, SNP sites with a depth of not less than 3X are selected to ensure the reliability of the data, and low-depth sites are filtered out;

[0060] (2) Integrity filtering: SNP markers with a genotype covering at least 50% of all sample individuals were screened; specifically, for each polymorphic marker locus, it was required that at least 50 samples out of 100 samples had a determined genotype, thereby filtering out markers with poor genotype integrity;

[0061] (3) Minor allele frequency (MAF) filtering: Sites with an MAF value lower than 0.05 were filtered out to ensure that the selected sites have sufficient variation information;

[0062] (4) After the above preliminary filtering, 40M basic SNP sites were retained.

[0063] (II) Deep filtering of SNPs

[0064] Linkage disequilibrium (LD)-based filtering: PLINK software was used for linkage disequilibrium analysis with two parameter settings: one was a 50-kb window, 10-bp step size, and R2 < 0.2; the other was a 100-kb window, 1-bp step size, and R2 < 0.2. These filtering parameters referred to the reported locus screening criteria for allopolyploid plants.

[0065] Hardy-Weinberg equilibrium (HWE) filtering: The p-values of the Hardy-Weinberg equilibrium for all SNP loci were calculated and filtered at 1e-6.

[0066] Gene region screening: SNP loci located in gene regions and their upstream and downstream regions were selected, and other loci within 100 bp upstream and downstream of the SNP loci were filtered out. For exon regions, if the distance between a certain SNP and the previous SNP was less than 100 bp and it was a non-synonymous mutation, then this SNP was retained.

[0067] Non-exon region filtering: In non-exon regions, variant loci of the A / T and G / C types were filtered out.

[0068] After deep filtering of SNPs, a total of 95,954 background loci, namely VIP SNP loci, were finally retained.

[0069] (III) Screening of trait-associated loci

[0070] In this example, using the resequencing data of 100 Leymus chinensis samples, a genome-wide association study (GWAS) was performed on 8 agronomic traits of Leymus chinensis using the mixed linear model (MLM) of GEMMA software. In the analysis, default parameters were used, and the significance threshold was set to 6 to identify the associated loci for each trait. The locus retention strategy was as follows:

[0071] (1) For traits with fewer than 450 associated loci, all associated loci were retained;

[0072] (2) For traits with more than 450 associated loci, loci less than 100K from the gene were screened, sorted by P-value, and the top 450 loci were selected;

[0073] (3) Among the screened loci, if the distance between two loci was less than 1K, only the locus closest to the gene was retained;

[0074] (4) Through these steps, a total of 1,667 loci associated with the agronomic traits of Leymus chinensis, namely VVIP SNP loci, were finally selected.

[0075] (IV) Determination of SNP loci on 50K liquid phase chips

[0076] After merging the SNPs of basic, VIP, and VVIP obtained in the above analysis and removing duplicates, the site dataset was screened using the sliding window method. In each window, SNPs of VIP and VVIP were preferentially selected, and the remaining positions were supplemented with background sites. 51,696 SNP sites can be screened out as the 50K chip site set, and the specific information of the core SNP sites is recorded in Table 1 below. Among them, the position and variation information of the SNP sites are represented in the form of chromosome: physical position: reference genotype / variant allele genotype, and the physical position of the SNP molecular marker combination is located and analyzed based on the Leymus chinensis reference genome.

[0077] Table 1 Core SNP Site Information

[0078]

[0079] Example 2

[0080] Based on the SNP site information screened in Example 1, the liquid chip probes were designed in this example.

[0081] The chip designed in this example is based on the principle of thermodynamic stability, including factors such as melting temperature (Tm), GC content, and sequence specificity, and comprehensively considers the genomic complexity of the target species, the position of the target site, and the GC content around it to ensure 99% capture of the conventional region. The probe is based on the positive strand of the Leymus chinensis reference genome, and the designed length range of the probe is 55 - 120 nt, with an average target length of 100 nt.

[0082] For complex regions with too high or too low GC content, by adjusting the probe position, a multi-layer 'tiling' probe design is adopted to improve the coverage of difficult regions. This scheme improves the capture efficiency in GC extreme regions by increasing the complementarity and stability between probes. In this example, the key design parameters of the 50K probe include:

[0083] (1)GC content: 30% - 80%; the number of sites near a single target site is less than 10; site specificity is greater than 50%; based on these criteria, the chip selects the set of sites with the lowest scores according to the scoring.

[0084] (2)1667 GWAS-related sites are retained without considering the scoring values.

[0085] Based on the above criteria, the liquid-phase chip probes were designed and synthesized in this example.

[0086] Example 3

[0087] In this example, the distribution and annotation of SNP sites on the chromosomes of the liquid-phase chip were carried out based on the facts.

[0088] In this example, the distribution of the liquid-phase chip screened in Example 1 on 14 chromosomes of Leymus chinensis was counted. On average, each chromosome contains about 3,689 SNPs, and the average distance between adjacent sites is 145 kb. The 51,697 SNPs screened showed an almost uniform distribution, and the site distribution map is as Figure 2 shown. The above SNP annotation results showed that among the 51,697 sites, 34% of the sites were located in the gene region and 2% were in the UTR region. In addition, 34% of the sites may have a significant impact on gene function, potentially leading to changes in the function of protein-coding genes. The annotation information of SNP sites is as Figure 3 shown. The SNP combination of Leymus chinensis screened by this chip significantly improved the detection success rate of candidate gene sites, providing strong support for subsequent functional gene research.

[0089] Example 4

[0090] In this example, based on the Leymus chinensis liquid-phase chip developed in the previous Example 2, 12 samples of Leymus chinensis were selected to test the performance of the chip, and the specific steps of the operation are as follows.

[0091] (I) Probe preparation

[0092] The preparation of the Leymus chinensis liquid-phase target capture probe described in this example includes template synthesis and RNA preparation. The probe template consists of single-stranded DNA (oligo) synthesized organically. After quality control (QC) and pooling, the qualified template is used to prepare RNA analog probes. In the synthesis of RNA probes, biotin-labeled nucleotide analogs (such as LNA, PNA, etc.) are used to improve the binding stability between the probe and the library. The biotin-labeled probe can bind to streptavidin-coated magnetic beads to further improve the capture efficiency.

[0093] (II) Library construction and capture

[0094] The DNA samples are sheared by ultrasound or enzyme digestion, end repair and 3' end A addition, adapter ligation and purification, and Pre-PCR library amplification to obtain the library required for hybrid capture. The capture method is carried out according to the SOP for hybrid capture, and after library and probe hybridization (12-18h), probe and magnetic bead binding, non-specific binding library rinsing, post-capture PCR amplification, library quantification and quality inspection, it is sequenced on the machine.

[0095] (III) Library quality control and quantification

[0096] Take 1 μL of library and use Qubit dsDNA HS Assay Kit reagent to detect the library concentration on Qubit 4.0 Fluorometer. The library concentration before capture >25 ng / μL is considered a qualified library.

[0097] Take 1 μL sample and use Qsep 100 for detection. The main peak of the library should be around 270-450 bp, with no impurity peaks before and after the main peak.

[0098] (IV) Sequencing on the machine

[0099] Take 1 μL of library and use Qubit dsDNA HS Assay Kit reagent to detect the library concentration on Qubit 4.0 Fluorometer and record the library concentration. The concentration of the library after capture is about 1-20 ng / μL.

[0100] Take 1 μL sample and use Qsep 100 to measure the length of library fragments. The length of the library is approximately between 270-450 bp.

[0101] Sequencing was performed using a high-throughput sequencing platform.

[0102] (V) Data analysis

[0103] In this embodiment, the raw image data obtained by high-throughput sequencing (Nova platform, MGI platform, etc.) is identified by a specific program, converted into raw sequencing data in Fastq format, and then compared with the reference genome for bioinformatics analysis, including the following steps:

[0104] (1) Quality control: Remove low-quality sequences, adapter sequences, ployG and other abnormal sequences in the original sequencing data to obtain clean data;

[0105] (2) Alignment: Align, sort, and remove duplicates of the Clean data with the reference genome to obtain a bam file;

[0106] (3) Statistics: Statistical analysis of indicators such as alignment rate, coverage rate, capture rate, and uniformity.

[0107] (VI) Test results

[0108] In this embodiment, after the above test analysis, the test results include various indicators of the target area, various indicators of the probe area, coverage of different areas, etc., which comprehensively and in detail demonstrate the various performances of the chip. The product has a high site detection rate, with an average site detection rate of 98.9%; the product has good stability and high genotype accuracy, and the average consistency rate of genotypes of repeated samples is 99%.

[0109] In this example, in order to intuitively display the expected performance and actual test performance of the chip, the QC rate, comparison rate, coverage, capture rate, uniformity and other indicators of two samples were taken to calculate the average value and draw a radar chart. The results are shown in the figure. Figure 4 As shown, it can be seen that all indicators are good and in line with expectations.

[0110] Example 5

[0111] This embodiment provides a liquid phase chip detection method for identifying Leymus chinensis varieties, using the Leymus chinensis 50K liquid phase chip prepared above to construct a library for sequencing Leymus chinensis variety samples, and the specific experimental process is as follows.

[0112] (I) Extraction and detection of DNA samples

[0113] In this embodiment, the samples to be tested, breed_1 to breed_11, are respectively: SZ1 Leymus chinensis strain, Nongmu No. 1 Leymus chinensis variety, Jisheng No. 4 Leymus chinensis variety, PL Leymus chinensis germplasm, Zhongke No. 3 Leymus chinensis variety, Jisheng No. 3 Leymus chinensis variety, Jisheng No. 2 Leymus chinensis variety, Zhongke No. 2 Leymus chinensis variety, Zhongke No. 1 Leymus chinensis variety, Jisheng No. 1 Leymus chinensis variety and Zhongke No. 5 Leymus chinensis variety. Among them, the sample size of Zhongke No. 5 Leymus chinensis variety (breed_11) is 30 individual plants, the sample size of Jisheng No. 3 Leymus chinensis variety (breed_6) and Jisheng No. 2 Leymus chinensis variety (breed_6) is 19 individual plants, and the sample size of the remaining varieties is 20 individual plants, for a total of 228 samples to be tested.

[0114] The leaves of the samples to be tested were taken, and the genomic DNA was extracted by CTAB method. The concentration, integrity and purity of the DNA were detected by agarose gel electrophoresis combined with Agilent5400.

[0115] (II) Library construction and sequencing

[0116] The 50K liquid phase chip technology of Leymus chinensis prepared by the above-mentioned embodiment comprises independently packaged 50K probe mixture and hybrid capture reagent. The probe mixture provides 50K liquid phase targeted capture probes in total.

[0117] As Figure 5 shown in the library construction process, it includes the following steps: The DNA sample is fragmented by sonication or enzymatic digestion, followed by end repair, 3'-end adenylation, adapter ligation and purification, and Pre-PCR library amplification, and finally a library that meets the requirements of hybridization capture is obtained. The capture method is strictly operated according to the SOP. The specific process includes: hybridization of the library with the probe (lasting 12 - 18 hours), binding of the probe to the magnetic beads, rinsing the non-specifically bound library, post-capture PCR amplification, library quantification and quality detection. After all steps are completed, the library is subjected to on-machine sequencing.

[0118] Example 6

[0119] In this example, bioinformatics standard analysis is performed on the data after sequencing in Example 5, specifically including data quality control and statistical analysis, SNP detection and annotation, and population genetic polymorphism analysis (population phylogenetic tree analysis, population principal component analysis, population structure analysis).

[0120] (I) Quality control analysis of sequencing data

[0121] The raw image data generated by high-throughput sequencing is identified and processed by a specific algorithm and converted into raw sequencing data in Fastq format (i.e., Raw data). After obtaining the raw data, it is aligned with the Leymus chinensis reference genome and subjected to bioinformatics analysis. The main analysis steps are as follows:

[0122] (1) Quality control: Quality assessment is performed on the raw sequencing data, and low-quality sequences, adapter sequences, and abnormal sequences such as polyG are removed to generate high-quality data (Clean data);

[0123] (2) Alignment processing: Align the Clean data to the reference genome, followed by sorting and deduplication to generate an alignment file (bam file);

[0124] (3) Statistical analysis: Statistical analysis is performed on key indicators such as alignment rate, coverage rate, capture rate, and uniformity. The results show that the average alignment rate of the sample is 93.76%.

[0125] (II) SNP detection and annotation

[0126] The samples were subjected to SNP detection using GATK 4.5 software, and preliminary filtration was performed using the VariantFiltration module. The default filtration conditions included: QD < 2.0, QUAL < 30.0, SOR > 3.0, FS > 60.0, MQ < 40.0, MQRankSum < -12.5, ReadPosRankSum < -8.0. On this basis, the following screening criteria were further adopted for in-depth filtration of high-quality SNPs: sequencing depth (Depth) < 3 for filtration, minor allele frequency (MAF) > 0.05, and missing rate (Miss) > 0.1. For SNP functional annotation, the efficient ANNOVAR tool was used to perform multi-dimensional analysis on the detection results by inputting the chromosomal position, start and end positions, reference base, and variant base information of the variant sites, including gene-based annotation, region-based annotation, filter-based annotation, etc. The SNP annotation results are as Figure 6 shown.

[0127] (III) Population phylogenetic tree analysis

[0128] After SNP detection, the obtained individual SNP data was used to calculate the genetic distance between populations. By using the TreeBeST software (Tree Building guided by Species Tree, http: / / treesoft.sourceforge.net / treebest.shtml), the distance matrix was calculated, and based on this, the neighbor-joining method (NJ method) was used to construct a phylogenetic tree. The bootstrap values of the phylogenetic tree were obtained by repeating the calculation 1000 times to ensure the reliability and statistical support of the results. The population phylogenetic tree analysis is as Figure 7 shown.

[0129] (IV) Population principal component analysis

[0130] Principal component analysis (PCA) is a commonly used data dimensionality reduction method that can be used to analyze population structure and clustering in genetics. In the present invention, principal component analysis (PCA) was used to analyze single nucleotide polymorphism (SNP) sites in genomic data, extract the main variables, and construct a population structure model. This method was further used to optimize the accuracy of genetic clustering analysis. The PCA analysis results are as Figure 8 shown.

[0131] (V) Population genetic structure analysis

[0132] Population structure analysis plays an important role in understanding the evolutionary process and population differentiation, and can further clarify the subpopulations to which individuals belong. Population structure analysis was carried out using the software AdmixStructure, and the analysis results are as follows Figure 9 shown. In the figure, each column represents an individual, and the length of different colored segments represents the proportion of the individual's genome derived from a specific ancestral population. The K values from 2 to 12 marked on the right side of the figure represent the range of the number of ancestral populations assumed in the analysis.

[0133] Example 7

[0134] In this example, advanced information analysis was continued for the sequencing data, and principal component analysis (PCA) was combined with the random forest algorithm for variety identification.

[0135] During the analysis process, first, the variety identification results were evaluated, and the key important loci were identified. Further evaluation was carried out on these loci to ensure their contribution to variety identification. Finally, combined with the accuracy evaluation of the model, a unique locus QR code for a specific variety was generated as an identifier for variety recognition.

[0136] (I) Evaluation of variety identification results

[0137] The SNP data obtained by sequencing was encoded, and the principal SNP loci were extracted by principal component analysis (PCA) and singular value decomposition (SVD). The scores of each SNP marker were calculated and ranked within the chromosome. Although a large number of SNP loci were selected after PCA preselection and feature filtering, in order to evaluate the importance of the selected loci, rules need to be established to assign samples to the corresponding varieties.

[0138] In this example, the random forest algorithm was used for modeling. The random forest realizes classification by integrating multiple decision trees, and its two key parameters are the number of decision trees (n_estimators) and the number of preselected variables at the nodes. n_estimators has a monotonic effect on the model accuracy, but when a specific threshold is reached, increasing n_estimators no longer significantly improves the accuracy, and the computational cost and training time increase accordingly. In the present invention, by balancing the model accuracy and the computational cost, the n_estimators when the accuracy tends to be stable was selected as the optimal parameter value.

[0139] The training set of each decision tree in the random forest is generated by the Bootstrap Sampling method, leaving 30% - 40% of the data unselected as "Out-of-Bag (OOB) data". The OOB data is used for evaluation during the generation process to obtain the OOB error rate or accuracy, providing an unbiased estimate of the classification effect. At the same time, to ensure the stability and reliability of the results, k-fold cross-validation is combined to evaluate the model.

[0140] (2) Obtaining important loci for variety identification

[0141] Use the trained model to score the OOB data and calculate the AUC or other evaluation metrics. Subsequently, the following operations are performed on each feature in the OOB data in turn:

[0142] (1) Randomly shuffle the values of the current feature;

[0143] (2) Re-score the data and calculate the evaluation metrics;

[0144] (3) Calculate the rate of change of the metrics. By calculating the rate of change for all features and sorting them, the importance of the features (i.e., the importance of the loci) is quantified.

[0145] (3) Evaluating important loci for variety identification

[0146] Sort the loci obtained during the variety identification process according to their importance and accumulate their importance values. When the accumulated importance reaches or exceeds 90%, select the corresponding locus set to retrain the random forest model. And use the trained model to evaluate the accuracy of the test data, and at the same time use the cross-validation method to verify the performance of the model.

[0147] (4) Model accuracy evaluation

[0148] To further verify the accuracy of the above model, the data of each population is randomly divided into a training set and a test set in a ratio of 7:3. Among them, the training set is used for model training, and the test set is used for model prediction. This process is repeated 100 times to verify the accuracy of the model in each prediction. The prediction results are as Figure 10 shown.

[0149] (5) QR code for variety-specific loci

[0150] Convert the identified SNP locus information into the corresponding QR code form. By scanning the QR code, the molecular identity information of the variety can be quickly obtained. The specific SNP fingerprint information of the variety is as Figure 11 shown.

[0151] In summary, the liquid chip detection method for Leymus chinensis variety identification according to the present invention is based on the 50K liquid chip of the whole genome of Leymus chinensis, determines the SNP molecular marker combination of the Leymus chinensis by using whole genome resequencing, screens out 51,696 SNP molecular marker sites based on the Leymus chinensis reference genome Leymus_chinensis_Lc6-5, develops the 50K liquid chip of the whole genome of Leymus chinensis based on the target capture sequencing technology, and realizes high-throughput and high-precision genotyping of the Leymus chinensis genome by combining high-density SNP site probes, and can comprehensively evaluate the gene characteristics of Leymus chinensis.

[0152] Obviously, the above embodiments are merely examples for clear illustration and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation manners here. And the obvious changes or modifications derived therefrom still fall within the protection scope of the present invention.

Claims

1. A liquid-phase chip detection method for the identification of Leymus chinensis varieties, characterized in that, The method includes the following steps: (1) Perform SNP detection on different varieties of Leymus chinensis based on the 50K liquid chip of the whole genome of Leymus chinensis. Among them, the liquid chip includes SNP sites located on the Leymus_chinensis_Lc6-5 reference genome of Leymus chinensis, covering gene regions and their related non-gene regions; The specific information of the core SNP sites of the liquid chip is shown in Table 1 of the specification; The 50K liquid chip of the whole genome of Leymus chinensis includes an independently packaged 50K probe mixture and reagents suitable for liquid hybridization capture; among them, The probe mixture contains 50K highly sequence-specific probes, which can identify and detect the core SNP molecular marker sites shown in Table 1 of the specification; The 50K liquid-phase targeted capture probe is designed based on the positive strand of the Leymus chinensis reference genome; The length range of the probe is 55-120 nt, and its average length is 100 nt; The 50K liquid-phase targeted capture probe adopts a 1X probe coverage in the conventional region and a multi-layer overlapping probe structure in the high-difficulty complex region; (2) Compare the SNP detection results with the Leymus chinensis reference genome, perform SNP analysis and annotation, and establish a SNP information database for different varieties.

2. The liquid-phase chip detection method for Leymus chinensis variety identification according to claim 1, wherein The analysis steps of the SNP include: using the SNPs identified by sequencing, calculating the distance matrix, and constructing a phylogenetic tree for different varieties.

3. The liquid chip detection method for Leymus chinensis variety identification according to claim 2, wherein, The method also includes the step of performing principal component analysis on the SNP data obtained by sequencing. Based on the degree of SNP difference in individual genomes, individuals are clustered into different subgroups according to different trait characteristics, and population genetic structure analysis is performed on the samples.

4. The liquid chip detection method for Leymus chinensis variety identification according to claim 3, characterized in that, The method also includes the step of encoding the SNP data set obtained by sequencing, and then extracting the main SNP sites by principal component analysis and singular value decomposition.

5. The liquid-phase chip detection method for Leymus chinensis variety identification according to claim 4, wherein, The method also includes the step of classifying the varieties of the samples using random forest; In the random forest processing, the optimal number of trees is determined after balancing accuracy and computational cost; In the random forest processing, the bootstrap method is used to generate decision trees, and the classification performance is evaluated through the out-of-bag (OOB) data that is not drawn, and an unbiased OOB error rate or accuracy rate is obtained; In the random forest processing, it also includes using k-fold cross-validation for model verification.

6. The liquid chip detection method for Leymus chinensis variety identification according to claim 5, wherein In the random forest processing, it includes: Using the trained model to score the OOB data, and calculating the AUC or other evaluation metrics; For each feature in the OOB data, perform the following operations in sequence: randomly shuffle the values of the feature to observe the change in model performance; re-score and calculate the evaluation metrics; calculate the change rate of the metrics; According to the importance ranking, accumulate the feature importance, select the sites whose accumulated value exceeds 90%, re-train the random forest model, and finally screen out the required number of SNP sites.

7. The liquid chip detection method for Leymus chinensis variety identification according to claim 6, wherein The annotation step of the SNP includes the step of encoding the screened SNP site information into a two-dimensional code for displaying the molecular identity information of the corresponding variety.

8. The liquid chip detection method for Leymus chinensis variety identification according to claim 7, wherein The method also includes the step of identifying the variety of the Leymus chinensis to be tested; it includes the step of obtaining the genotype data of the individual Leymus chinensis to be tested based on the 50K liquid-phase chip of the whole genome of Leymus chinensis, and the step of comparing and identifying the Leymus chinensis variety based on the constructed reference sample library.

Images