Whole-genome low-density chip for breeding high-yield and long-life characters of dairy cow and application of whole-genome low-density chip

Abstract

The invention discloses a whole-genome low-density chip for breeding high-yield and long-life characters of dairy cows and application of the whole-genome low-density chip. The invention firstly discloses a whole-genome low-density chip for breeding the high-yield and long-life character of the dairy cow, the whole-genome low-density chip comprises 10000 SNP (Single Nucleotide Polymorphism) loci, ARS-UCD1.2 is used as a reference genome, and the information of the 10000 SNP loci is shown in a table 6. The invention further discloses application of the whole-genome low-density chip. The 10,000 SNP loci contained in the whole genome low-density chip are uniformly distributed on chromosomes, cover the whole genome and important character loci, can effectively improve the longevity and the accuracy of milk yield breeding, and can be applied to gene screening, gene positioning, marker-assisted breeding and the like. In addition, the detection cost of the low-density chip is greatly reduced, and the low-density chip is more suitable for commercialized dairy farms and has important significance on dairy cow breeding and dairy cow group improvement.

Description

Technical Field

[0001] The present invention relates to the fields of molecular biology and biochips, and more specifically, to a whole genome low-density chip for breeding high-yield and long-lived dairy cows and its application. Background Art

[0002] Since the concept of genomic selection was proposed in 2001, genomic selection has become a conventional breeding tool for dairy cattle breeding. Genomic selection uses markers covering the entire genome and can achieve early selection of newborn offspring through linkage disequilibrium between markers and trait-related QTLs.

[0003] The implementation of genomic selection requires a genomic breeding chip of the corresponding variety. To determine the genetic markers in the chip, it is first necessary to determine the genetic loci associated with the target trait. Common methods for mining trait-related loci include Genome-Wide Association Studies (GWAS) and Fixation index (Fst).

[0004] Genome-wide association analysis refers to the study of associations between genes and phenotypes at the genome-wide level with large samples and repeated verification. It is a research method that searches for genetic factors related to complex traits by typing large-scale population DNA samples with high-density genetic markers (such as SNPs or CNVs) throughout the genome, and comprehensively reveals genetic genes related to phenotypes. Single nucleotide polymorphism (SNP) mainly refers to DNA sequence polymorphisms caused by variations in a single nucleotide at the genome level.

[0005] The genetic differentiation index between populations is a measure of population differentiation and genetic distance. The larger the differentiation index, the greater the difference. It is suitable for comparing the diversity between subpopulations. It is used to measure the degree of population differentiation. The value ranges from 0 to 1. 0 means that the two populations are randomly mated and the genotypes are completely similar; 1 means that they are completely isolated and completely dissimilar. It is often estimated from the diversity of genes, such as SNP. It is a population genetics statistical method based on the Harwin equilibrium.

[0006] SNP analysis is currently a common method for mining key genes. However, China's dairy cow breeding work (frozen semen, embryos, genetic evaluation) all relies on foreign countries, especially genetic evaluation work requires the collection of blood, hair and other samples to be sent to the United States for testing, which has problems such as leakage of dairy cow genetic resources, long testing cycle and high testing cost; more than 70% of frozen semen relies on foreign imports, and the independent breeding capacity is relatively low.

[0007] Saikexing Group has 26 ranches with 152,500 cattle in stock. As of May 2024, the group's herd yield has exceeded 40kg. Among them, 16 ranches have been carrying out DHI work for a long time, with a total of 98,300 cattle; there are 8 ranches where DHI work was carried out earlier and data records are complete, with a total of 59,300 cattle tested. There is sufficient data to carry out analysis of traits such as longevity and milk production. In addition, Saikexing Group also has 2 bull stations and 2 national core breeding farms. A total of more than 440 excellent bulls have been bred, and the production capacity of sex-controlled frozen semen is the first in Asia and the second in the world. The annual production of various frozen semen is about 3 million doses, and the cumulative promotion and application nationwide has exceeded 7.4 million doses. The sales volume of frozen semen products has ranked first in the country for twelve consecutive years, and it is the breeding company with the highest market share in China. Therefore, the company has genetic relationship data of offspring of different bulls in different regions and different ranches.

[0008] Longevity refers to the overall production time of dairy cows and is used to quantify the potential of dairy cows to bring benefits. Longevity is usually measured in different indicators. For example, the production life span refers to the time from the first calving to the death or culling of dairy cows. Although the production life span does not affect the unit yield, increasing the production life span is conducive to reducing the cost of passive culling and purchasing cows, and can also increase lifetime milk production. It is an important economic indicator for pasture production and management; for example, the number of parities is used to indicate the time the cow is in the group and the time it produces milk. Production performance mainly includes: milk yield, milk fat rate, and milk protein rate. This trait is controlled by a large number of micro-effect polygenes, affected by genetic and environmental factors, and is a quantitative genetic trait. Traditional phenotype-based breeding methods are difficult to achieve significant results in a short period of time. In addition, due to the micro-effect polygenes, the results of molecular-assisted selection marker breeding methods are not ideal.

[0009] The longevity and milk production of dairy cows are usually due to the superposition of genetic and environmental factors. Therefore, under the condition that environmental factors are relatively fixed, it is necessary to analyze and explore the genetic information such as genetic loci and genes that affect high-yield and long-lived dairy cows. Developing breeding chips to identify high-yield and long-lived dairy cows is of great significance for dairy cow breeding. Summary of the invention

[0010] One object of the present invention is to provide a whole genome low-density chip for the selection of high-yield and long-lived traits of dairy cows, so as to improve the accuracy of evaluating traits such as longevity and milk production of a herd and increase benefits.

[0011] Another object of the present invention is to provide an application of the above-mentioned whole genome low-density chip.

[0012] In order to achieve the above object, the present invention adopts the following technical solutions:

[0013] The present invention first provides a whole genome low-density chip for breeding high-yield and longevity traits of dairy cows, wherein the whole genome low-density chip comprises 10,000 SNP sites, and ARS-UCD1.2 is used as a reference genome. The information of the 10,000 SNP sites is shown in Table 6.

[0014] Furthermore, the SNP site is associated with the longevity and milk production traits of dairy cows.

[0015] The present invention further provides the use of the above-mentioned cow whole genome low-density chip in any of the following:

[0016] 1) Application in dairy cow breeding;

[0017] 2) Application in genome-wide association analysis of dairy cows;

[0018] 3) Application in the breeding of longevity traits in dairy cows;

[0019] 4) Application in the breeding of milk production traits in dairy cows.

[0020] In the present invention, the longevity is measured by the number of parities. Individuals with more than 5 parities are defined as long-lived or high-reproduction individuals, and those with less than 3 parities are defined as non-long-lived or low-reproduction individuals. The milk production is measured by the mean of the DHI data, with an average milk production of 40kg as the threshold. Cows with a milk production above 40kg are defined as high-producing cows, and those with a milk production below 40kg are defined as low-producing cows.

[0021] The present invention further provides a method for preparing a whole genome low-density chip for breeding high-yield and long-lived traits of dairy cows, wherein the preparation method comprises forming the above 10,000 SNP sites into a low-density chip.

[0022] In the present invention, the dairy cow is a Holstein cow.

[0023] The beneficial effects of the present invention are as follows:

[0024] The whole genome low-density chip for high-yield and long-lived traits of dairy cows of the present invention contains 10,000 SNP sites, which are evenly distributed on chromosomes, cover the whole genome, and include important trait sites, and can effectively improve the accuracy of longevity and milk production trait selection, and can be applied to gene screening, gene positioning, marker-assisted breeding, etc. In addition, the detection cost of the low-density chip of the present invention is greatly reduced, and it is more suitable for use in commercial dairy farms, which is of great significance to dairy cow breeding and dairy cow population improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0026] Figure 1 Customize the flow chart for the probe.

[0027] Figure 2 It is the site-chromosome distribution map, where the horizontal axis is the chromosome and the vertical axis is the number of sites.

[0028] Figure 3 This is the marker distribution density map on the chromosome.

[0029] Figure 4 It is a MAF distribution diagram, where the horizontal axis MAF is the different ranges of MAF values, and the vertical axis Number is the number of core sites within the MAF range. The higher the MAF value of the target site, the better the polymorphism of the site.

[0030] Figure 5 This is a statistical chart of SNP variation types.

[0031] Figure 6 It is a functional site distribution map, where the horizontal axis Type is the gene structure type, and the vertical axis Number is the number of target sites under the relevant type.

[0032] Figure 7 The PCA visualization results are shown in Figure 2, where the upper figure is the PCA obtained based on the original genotype of the population, and the lower figure is the PCA obtained based on the chip SNP sites.

[0033] Figure 8 A graph showing the predictive effects of different machine learning methods on milk production (A) and longevity (B) for the 10K chip. DETAILED DESCRIPTION

[0034] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the content specifically described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0035] Saikexing currently has 6,932 long-lived cattle with ≥5 parities in the group, accounting for 4.54% of the group size. The present invention selected 600 of the 963 cattle born after August 7, 2004, with clear three-generation pedigrees and 5 to 11 parities (from Saikexing's 8 ranches), which are representative. In addition, the above cattle have been subjected to DHI measurement since July 13, 2016, and the measurement time is relatively early. After data cleaning, the cumulative number of participating cattle is 59,300, and the data volume is large, which ensures the accuracy of phenotypic data and genome selection. Therefore, Saikexing Company has the conditions to carry out genome-wide association analysis for dairy cows, explore functional sites, develop breeding chips, and improve the accuracy of longevity and milk production trait selection to improve efficiency.

[0036] Example 1 Design and preparation of low-density microarray of the whole genome of dairy cows

[0037] 1. Sample collection

[0038] The samples were collected from 600 Chinese Holstein cows in 8 ranches of Saikexing, and all individuals had phenotypic and genotypic data. Among them, the phenotypic data were divided into DHI data and ranch management data.

[0039] 2. Genome Resequencing

[0040] 1. Genomic DNA extraction (magnetic bead method)

[0041] Preparation for DNA extraction: Dissolve Proteinase K in Proteinase K Storage Buffer to prepare 20 mg / ml Proteinase K solution and store at -20℃ for later use.

[0042] Extraction steps:

[0043] (1) Add the corresponding reagents to the 96DW deep well plate according to Table 1:

[0044] Table 1 DNA advance reagent addition position

[0045]

[0046] (2) Add blood samples to the corresponding wells of Plate 1, with a sample volume of 300 μL / sample.

[0047] (3) Insert the Spin tips Pack into the 96DW deep well plate.

[0048] (4) Run the blood DNA extraction program. According to the instrument program flow, place the deep-well plate with the magnetic rod cover and the deep-well plate with reagents into the designated positions of the instrument.

[0049] (5) Obtain genomic DNA (gDNA).

[0050] 2. Genomic DNA quality inspection

[0051] (1) Detection of DNA integrity: Take the gDNA to be detected, spot it on 1.5% agarose gel, and perform electrophoresis at 150 V for 25 min.

[0052] (2) DNA purity detection: The purity of DNA was detected using NanoDrop 2000 nucleic acid protein analyzer. The A260 / A280 ratio was in the range of 1.7-2.1, and the A260 / A230 ratio was in the range of 1.8-2.2, which met the requirements for library construction.

[0053] (3) Quantification of DNA concentration: gDNA was quantified using the dsDNAHS Assay Kit for Qubit.

[0054] 3. Library construction

[0055] The specific steps of library construction are as follows: Detect qualified genomic DNA and use a universal library construction kit (forMGI) Library construction, DNA was crushed by Covaris TM Randomly interrupted, end-repaired, A-tailed, and connected to sequencing adapters, SP Beads screened out fragments of about 300-350 bp, amplified them by PCR, and used SPBeads were used to purify the PCR products and finally obtain the sequencing library.

[0056] 4. Hybridization capture

[0057] After quantification by Qubit2.0 and (Agilent) detects the insert fragments of the sequencing library. Library hybridization: After the library is qualified, the biotin-labeled probe is specifically combined with the DNA library with the adapter sequence. The hybridization system consists of probes, hybridization buffer, blocking reagents and DNA libraries. The entire hybridization reaction is carried out on a PCR instrument, requiring precise temperature control to prevent the system from evaporating. Hybridization capture: After the hybridization reaction is completed, streptavidin magnetic beads are added to the reaction system for hybridization capture. Streptavidin magnetic beads adsorb the double-stranded complex of the biotin-labeled probe and DNA, and the magnetic beads are adsorbed by a magnetic stand to obtain a hybrid capture library. Library product elution: After the capture is completed, the hybrid capture library is eluted with an elution solution to remove non-specific hybridization. High temperature and low salt environment treatment: In a high temperature and low salt environment, a large number of non-specific hybridization complexes cannot exist stably and can be removed by multiple washings. Capture product enrichment: Through post-PCR, the eluted product is amplified and the library is removed. Reasonably adjust the number of amplification cycles according to the size of the liquid phase probe.

[0058] 5. Sequencing

[0059] After the library inspection is qualified, pooling is performed according to the effective concentration of the library and the target amount of data to be downloaded from the machine. The DNBSEQ-T7 sequencer is used to select the PE150 sequencing strategy for sequencing.

[0060] The principle of sequencing is to use DNA nanoball (DNB) core sequencing technology. First, DNA molecular anchors and fluorescent probes are polymerized on the nanoballs, and then the high-resolution imaging system collects the light signal. After the light signal is digitally processed, the sequence to be tested can be obtained. The raw image data obtained by sequencing is converted into raw sequence data (rawreads) by base recognition software (Base Calling), and the data is stored in FASTQ file format.

[0061] 3. Trait association site mining

[0062] 1. Low-depth sequencing filling

[0063] The low-depth sequencing population was filled to a high-depth level. Specifically, 1) a filling reference panel was constructed. The cattle sequencing data was downloaded from the public database. At the same time, the cattle in this population were sequenced at a high depth. The genotype data obtained after SNP calling was used to construct a filling reference panel, and a total of 45,008,371 SNPs were obtained; 2) the sequencing data of the low-depth sequencing population was aligned and quality controlled to obtain the reads file (bam) format, and the low-depth sequencing population was filled to a high-depth level using GLIMPSE2 software. 3) The filled population was quality controlled, and sites with MAF < 0.01 and significant deviation from the Hartmann equilibrium were eliminated, leaving 9,927,289 SNPs for subsequent analysis.

[0064] 2. Site screening

[0065] The traits tested this time are divided into two categories: longevity and milk production. Regarding the definition of traits, longevity is measured by parity. Individuals with more than 5 parities are long-lived or high-fertility individuals, and those with less than 3 parities are defined as non-long-lived or low-fertility individuals; milk production is measured by the mean of DHI data, with an average milk production of 40kg as the threshold. Cows with a milk production above 40kg are defined as high-yield cows, and cows with a milk production below 40kg are defined as low-yield cows.

[0066] 2.1. GWAS analysis of binary traits

[0067] The GEMMA software was used to perform genome-wide association analysis based on a logit regression model for binary traits. The phenotype yi of the i-th individual can be defined as yi = 1 if individual i belongs to the case, and yi = 1 if individual i belongs to the control. The model for the k-th variant was constructed as follows:

[0068] Logit(π)=Xα+G k β+g,

[0069] Where, π=P(y=1|X,G k , g) is a given covariate (X) and genotype (G k) and the random effect g, the probability vector of an individual being defined as high-yield (i.e., average milk production greater than 40 kg) or long-lived (more than 5 parities).

[0070] 2.2. Mining of population differentiation sites

[0071] In population differentiation, the Fst method is often used to identify regions or sites with population differences. This method uses the Weir and Cockerham method in vcftools to estimate the Fst score of genetic markers and calculate the fixation index (Fst) to determine the priority of SNPs. The value range of Fst is 0 to 1. The larger the value, the greater the degree of differentiation between populations. When the Fst index is 1, it indicates that the genetic variation is completely differentiated in each population; when the Fst index is less than 0, there is no specific biological explanation and it can be ignored; when the F.-index is 0 to 0.05, it indicates that the degree of differentiation between populations is very small; when the F-index is 0.05 to 0.25, it indicates that there is a moderate degree of differentiation between populations; when the F-index is greater than 0.25, it indicates that the degree of differentiation between populations is very large.

[0072] Based on the above GWAS and population differentiation analysis, 1495 loci and 2889 SNPs related to milk production and longevity were obtained.

[0073] 4. Chip Design

[0074] Based on the trait-associated loci obtained from the above loci mining, the genetic loci in different regions of the genome are further supplemented according to the following genetic locus selection principles (1 to 7 below) to reach the 10K chip locus level. The detailed technical points are as follows:

[0075] 1. Site screening principles

[0076] For the candidate loci to be added to the chip, that is, the genetic loci on the genome that need to be added to the 10K chip, screening is performed considering their characteristics on the genome:

[0077] (1) Covers relevant loci, such as QTL, GWAS, domestication-related genes, inter-species and intra-species subpopulation differences, excellent traits, terminators, alternative splicing, non-synonymous mutations, etc.

[0078] (2) High polymorphism (MAF ≥ 0.05 by default).

[0079] (3) Uniform coverage of the entire genome.

[0080] (4) Multiple site selection: Generally, 2-4 times the target site is selected as the candidate site set.

[0081] (5) The design system scores and determines the final chip customization site.

[0082] 2. Site selection priority

[0083] Based on the key points of "1. Site screening principles", the genetic sites on the genome used to make the 10K chip were preliminarily screened. Since there may be multiple candidate sites in a region, the importance levels are set considering their different functions, and the candidate sites obtained based on the previous screening principles are ranked in importance, and then the sites are selected to ensure that the sites in the 10K chip are evenly distributed on the genome:

[0084] (1) Priority 1 = VIP loci, QTL loci, GWAS loci, important genes, common genes, etc.;

[0085] (2) Priority 2 = terminators on annotated genes (select all, but refer to MAF values) / variable splicing (select all, but refer to MAF values) / non-synonymous mutation sites (select all, but refer to MAF values), etc.;

[0086] (3) Priority 3 = whole genome coverage sites.

[0087] Based on the above importance ranking principles, sites in a region are selected to determine candidate sites in the chip.

[0088] 3. Probe customization process

[0089] Probe customization flow chart Figure 1 shown.

[0090] 4. Chip customization information is shown in Table 2.

[0091] Table 2 Chip customization information

[0092]

[0093] 5. Site quality control

[0094] Based on the above strategies of steps 1-4, the genomic genetic variation sites of the low-depth sequencing population were typed and checked, and data quality control was performed to obtain accurate and high-quality variation sites. The quality control content is shown in Table 3 (i.e., 9927289 SNP sites after filling and quality control in step 3-1 low-depth sequencing), including: reads coverage sequencing depth, deletion rate, kinship detection, minimum allele frequency (MAF), Hardy-Weinberg equilibrium, outlier samples, etc. That is, after deleting low-quality samples and sites, 370 samples remained (as shown in Table 4), and 9927289 SNP sites in the genome were used as candidate site sets for subsequent chip site selection.

[0095] Table 3 Quality control parameters and thresholds

[0096] # Quality control content Reference threshold 1 Harz balance test 0.000001 2 Relatedness test (animals) 0.25 3 Minimum allele frequency (MAF) 0.05 4 Genotype missing rate 0.1 5 Missing sample rate 0.2 6 Site read coverage depth assessment 1 / 2 of the average depth 7 Outlier Samples Filter by PCA results

[0097] Table 4 Sample details

[0098] Variety strain Original number of samples The remaining number of samples Group1 600 370 total 600 370

[0099] 6. Site Scoring

[0100] The candidate site set to be selected is submitted to a special probe design system for scoring. The optimal probe is determined based on the evaluation results of the upstream and downstream sequences of the target site. The designed probe is used for subsequent sequencing on the machine to capture the 10K chip site for the target sequencing individual. The specificity, complexity, GC content, etc. of the upstream and downstream sequences of the target site are mainly evaluated. The target site is placed in the middle of the probe first, and the designed probe is 120bp in length.

[0101] According to the scoring results, a total of 20K high-quality chip customized sites were selected as the final target site set of the chip, namely the 1x site set. Then, sites were selected from the 1x site set (20K) through site evaluation and cyclic testing to obtain the optimal 10K site set, namely the 10K chip.

[0102] 7. Site evaluation

[0103] The loci set (10K loci) for the proposed customized chip was selected from the 1x loci set (20K), and the loci set (10K loci) for the proposed customized chip was evaluated, mainly including the loci set ranking, chromosome distribution of loci, loci density distribution, MAF distribution, mutation spectrum analysis, marker function loci distribution statistics, PCA population structure and other indicators of the final chip product. The site evaluation results show that the sites evenly cover the whole genome, covering important trait sites, and can be applied to gene screening, gene positioning, marker-assisted breeding and other directions.

[0104] The specific method is as follows:

[0105] 7.1 Site set ranking

[0106] According to the distribution of the functional intervals of the screened sites on the genome, the sites are ranked and counted according to their importance or priority. See Table 5 for details.

[0107] Table 5 Site set ranking table

[0108] Tilingorder Number of sites Proportion 1 2272 22.72％ 2 3897 38.97％ 3 3831 38.31％ SUM 10000 100％

[0109] Tiling order 1 = VIP, QTL loci, GWAS loci, important genes, common genes, etc.;

[0110] tiling order2 = annotation of promoter, terminator, alternative splicing, etc.;

[0111] Tiling order 3 = whole genome coverage of sites.

[0112] 7.2 Loci chromosome distribution

[0113] The selected points are plotted using the R package, and the results are as follows Figure 2 From the figure, we can see the priority distribution of the variant sites on each chromosome. Therefore, the distribution of the priority sites on the chromosome of this chip is relatively uniform, and the coverage is good.

[0114] 7.3 Site density distribution

[0115] Figure 3 This is the distribution density map of the chip loci set on each chromosome. From the figure, we can see that the loci of each chromosome are distributed evenly, indicating that the loci contained in the chip can cover each chromosome well.

[0116] 7.4 MAF Distribution

[0117] MAF is the minimum allele frequency, which usually refers to the frequency of uncommon alleles in a given population. For example, for the three genotypes TT, TC, and CC, the frequency of C in the population is 0.36, and the frequency of T is 0.64. Then allele C is the minimum allele, and the minimum allele frequency MAF is 0.36. Figure 4 This is the MAF distribution diagram. It can be seen from the figure that the sites with the minimum allele frequency > 0.1 account for 99.99%.

[0118] 7.5 Mutation Spectrum Analysis

[0119] Base transition refers to the replacement between purine and purine, and between pyrimidine and pyrimidine; base transversion refers to the replacement between purine and pyrimidine.

[0120] Due to base structure reasons, the probability of conversion is usually higher than that of transversion. After detecting and filtering high-confidence SNPs, the SNP mutation spectrum is statistically analyzed. Point mutations include 6 types: A / T, A / C, A / G, G / C, G / T, G / A. For example, A / C represents the mutation type from A to C on a chain. By classifying and counting the mutation results of all samples, the percentage of each mutation type in each sample can be seen. Figure 5 This is a statistical chart of SNP variation types. It can be seen from the chart that A / G and C / T mutation types account for a high proportion, and other variation types are similar.

[0121] 7.6 Distribution statistics of marker functional sites

[0122] ANNOVAR software was used to annotate the site set (10K sites) of the planned customized chip, and the 10K sites were aligned to the corresponding regions on the genome. Figure 6 This is a functional site distribution map. It can be seen from the figure that there are more non-synonymous mutations, intergenic regions and SNPs in the screened sites.

[0123] 7.7 PCA Population Structure

[0124] PCA analysis uses a graph to show the clustering of individuals within a population. The results of PCA analysis can provide a general understanding of the genetic structure of a population. On a PCA graph, individuals with closer linear distances in space usually have closer genetic relationships, while individuals with farther linear distances in space have farther genetic relationships. This can provide good identification results for chips containing variety identification sites. Figure 7 This is a PCA diagram. It can be seen from the figure that the sample clustering of the PCA obtained from the original genotype of the population and the PCA obtained from the chip SNP sites are consistent in both, indicating that the chip can better represent the original status of the population.

[0125] 8. Chip design results

[0126] Based on the 600 samples provided, a total of 10K SNP loci were finally obtained through chip design condition screening and uniform distribution considerations (the specific SNP loci information is shown in Table 6), and a 10K chip was successfully designed for breeding dairy cows for longevity and milk production traits.

[0127] Table 6 Information on SNP loci in the microarray for longevity and milk production traits in dairy cows

[0128]

[0129]

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142]

[0143]

[0144]

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165]

[0166] Example 2 Chip effect verification test

[0167] A total of 600 cattle were tested to verify the effect of the 10K chip obtained in Example 1. The phenotype was predicted using a machine learning method by repeating the 5-fold cross-validation strategy 20 times to verify the application effect of the 10K chip.

[0168] This test used four machine learning methods: support vector machine (SVM), random forest (RF), k-nearest neighbor (KNN) and naive Bayes (NB) machine learning methods to predict milk production and longevity traits. Longevity is measured by parity, with more than 5 parities defined as longevity and less than 3 parities defined as non-longevity. Milk production is measured by the mean of DHI data, with an average milk production of 40kg as the threshold. Cows above 40kg are defined as high-yielding cows and cows below 40kg are defined as low-yielding cows. Therefore, milk production and longevity are predicted as binary traits, and the prediction accuracy measurement indicator is expressed as the proportion of cows that are accurately judged as high-yielding or low-yielding, long-lived or non-long-lived, that is, TP / (TP+FP). TP is the number of accurately classified cows, and FP is the number of incorrectly classified cows. The results are shown in Tables 7 and Figure 8 This indicates that the 10K chip obtained in Example 1 has a good predictive effect on milk production and longevity traits.

[0169] Table 7 Prediction effect of 10K chip on milk production and longevity

[0170]

[0171] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the protection scope of the present invention.

Claims

1. A whole genome low-density chip for breeding high-yield and long-lived dairy cows, characterized in that: The whole genome low-density chip contains 10,000 SNP sites, with ARS-UCD1.2 as the reference genome. The information of the 10,000 SNP sites is shown in Table 6.

2. The whole genome low-density chip according to claim 1, characterized in that: The SNP loci are associated with longevity and milk production traits of dairy cows.

3. The cow whole genome low-density chip according to claim 1 or 2, characterized in that: The dairy cows are Holstein cows.

4. Use of the whole genome low-density chip according to any one of claims 1 to 3 in dairy cow breeding.

5. Use of the whole genome low-density chip according to any one of claims 1 to 3 in whole genome association analysis of dairy cows.

6. Use of the whole genome low-density chip according to any one of claims 1 to 3 in the breeding of longevity traits in dairy cows.

7. Use of the whole genome low-density chip according to any one of claims 1 to 3 in the breeding of dairy cows for milk production traits.

8. The use according to any one of claims 4 to 7, characterized in that: The dairy cows are Holstein cows.

9. The method for preparing a whole genome low-density chip according to any one of claims 1 to 3, characterized in that: The preparation method is to form a low-density chip with the 10,000 SNP sites described in any one of claims 1 to 3.

10. The preparation method according to claim 9, characterized in that: The dairy cows are Holstein cows.

Images