SNP (Single Nucleotide Polymorphism) combination for identifying Baicheng fatty chicken variety and identification method

Abstract

According to the SNP combination for identifying the Baicheng fatty chicken variety and the identification method, the SNP combination for identifying the Baicheng fatty chicken variety is a combination of 200 SNP sites, whether an unknown chicken variety is the Baicheng fatty chicken or not is identified by sequencing the SNP combination and combining with a machine learning model, the Baicheng fatty chicken can be distinguished from other varieties with relatively high accuracy, and the identification accuracy of the Baicheng fatty chicken variety is improved. The difficulty and time of variety classification are greatly reduced, and a new method is provided for Baicheng fatty chicken variety identification work. The method is of great significance to protection and utilization of germplasm resources of Baicheng fatty chickens.

Description

Technical Field

[0001] The present application belongs to the field of poultry breed identification. Specifically, the present application provides a SNP combination and identification method for Baicheng oily chicken breed identification. Background Art

[0002] The Baicheng Oily Chicken is named for its origins in the Baicheng Basin of Aksu Prefecture, Xinjiang, and for the distinctive, fragrant layer of fat deposited on its breast and other areas. A dual-purpose breed for both meat and eggs, the Baicheng Oily Chicken boasts a plump, thin-boned, tender, and richly flavored meat, earning it a high reputation among consumers for its nutritional value. However, Baicheng Oily Chickens have a poor reproductive performance. In the early days, they were primarily produced and sold on a small scale by farmers in the production area. Later, large-scale breeding programs introduced a large number of other breeds, resulting in significant crossbreeding. On January 15, 2010, the Baicheng Oily Chicken was listed in the National List of Livestock and Poultry Genetic Resources.

[0003] Machine learning is an emerging technology that uses computer systems to simulate human learning. It focuses on automatically acquiring knowledge and skills through algorithms and data to continuously improve model performance. Machine learning is a fundamental tool for understanding and exploring data science problems and a crucial tool in fields such as data analysis. It holds strategic significance for the development of cultivar classification. Machine learning differs from traditional data analysis techniques. Traditional methods primarily focus on processing data based on preset rules to ensure that results conform to specific expected patterns. Machine learning, on the other hand, emphasizes a data-driven approach, enabling models to self-learn and adapt, thereby improving their predictive capabilities and achieving accurate classification and prediction. Machine learning can effectively solve classification and prediction problems, which is of great significance to cultivar identification and data mining. Compared to traditional methods, machine learning can better capture data characteristics and build flexible models for classification and prediction. Traditional methods struggle to fully utilize all data features, resulting in relatively rigid and limited classification and prediction methods.

[0004] A commonly used machine learning algorithm is the support vector machine (SVM). SVMs are a binary classification model whose basic model is a linear classifier with the largest margin defined in feature space. They are one of the most widely used machine learning algorithms. SVMs perform well on both high- and low-dimensional data. They also boast robustness, strong generalization capabilities, and a relatively intuitive model. Due to their high efficiency and strong generalization, SVMs have been widely used in variety classification.

[0005] The basic principle of a support vector machine can be simply summarized as: it separates data from different categories by finding an optimal decision boundary and maximizing the "margin" between categories. This maximization of the margin distinguishes it from a perceptron; the SVM also incorporates kernel techniques, making it a truly nonlinear classifier. In breed classification, it can learn the SNP characteristics of different breeds and use this information to build a model to complete the classification task.

[0006] In recent years, with the rapid development of sequencing technology and machine learning, the conservation and utilization of germplasm resources has become a hot topic. With increasing global communication, the exchange of kinship between livestock and poultry breeds is also increasing, making the protection and utilization of genetic resources both a key and challenging aspect of germplasm conservation and utilization. Sequencing technology can easily capture a large number of single-nucleotide polymorphisms (SNPs). Simultaneously, by analyzing the information content of SNPs using population genetic parameters and identifying high-information SNP loci for analysis, it is possible to effectively perform breed classification. To better protect and utilize these germplasm resources, it is necessary to establish a comprehensive, feasible, and highly accurate breed allocation process and method as soon as possible.

[0007] Currently, machine learning algorithms are widely used due to their high accuracy and strong generalization. Several related methods have been developed for cattle, sheep, and pigs, using different population genetic parameters to screen for high-information SNP sites and using machine learning methods to build models and predict breed classification. However, there is currently little research on chicken breed distribution. The molecular information of many chicken breeds is unclear, and identification methods are relatively simple, resulting in insufficient identification and breeding technology for new breeds. We have utilized sequencing technology, population genetic parameter analysis, and machine learning algorithms to establish a process and model for Baicheng Oily Chicken breed identification and validated the feasibility of this method. This method can provide technical support for the subsequent classification and protection of other chicken breeds and play an important role in promoting the comprehensive revitalization of the seed industry. Summary of the Invention

[0008] On the one hand, the present application provides a SNP combination for identifying the Baicheng oily chicken variety, and the SNP combination includes the SNP sites shown in Table 1 of the specification.

[0009] On the other hand, the present application provides a method for identifying Baicheng oily chicken varieties, which includes using a machine learning model to process the SNP site information shown in Table 1 of the chicken sample to be tested.

[0010] Furthermore, the method comprises:

[0011] (1) Extracting genomic DNA from the chicken sample to be tested;

[0012] (2) detecting the genotype of the SNP site shown in Table 1 in the genomic DNA obtained in step (1) to obtain the genotype information of the SNP site;

[0013] (3) Use a machine learning model based on supervised learning to process the SNP site genotype information obtained in step (2).

[0014] Furthermore, the machine learning model based on supervised learning is a linear kernel support vector machine model.

[0015] Furthermore, in step (2), whole genome resequencing, SNP gene chip method, mass spectrometry, SNaPshot method, KASP method, ARMS method, CAPS method, molecular beacon method, melting curve method, Taqman probe or direct sequencing method are used to detect the genotype of the SNP sites shown in Table 1 in the genomic DNA obtained in step (1).

[0016] Since the number of SNP sites reaches more than 200, it is further preferred to use whole genome resequencing and SNP gene chip methods to detect the genotypes of the SNP sites shown in Table 1 in the genomic DNA obtained in step (1).

[0017] Furthermore, if there is any missing site information in the SNP site genotype information obtained in step (2), random filling is performed to fill it.

[0018] On the other hand, the present application provides the use of a gene chip for detecting the above-mentioned SNP combination in the preparation of a Baicheng oily chicken variety identification kit.

[0019] On the other hand, the present application provides a method for constructing a Baicheng oily chicken variety identification model based on machine learning, characterized in that the method comprises:

[0020] (A) Obtain SNP data for various chicken breeds, including Baicheng Oily Chicken, and perform quality control;

[0021] (B) Calculate genetic parameters for the quality-controlled SNP data obtained in (A), calculate the Euclidean genetic distance of the quality-controlled SNP data, and rank them accordingly;

[0022] (C) Some SNP sites with higher European genetic distance were selected to construct the Baicheng oily chicken breed identification model based on machine learning.

[0023] Furthermore, (A) obtains SNP data of different duck breeds by one or more methods selected from database download, whole genome resequencing sequencing, and SNP gene chip sequencing; obtains SNP data of different duck breeds by whole genome resequencing sequencing or gene chip sequencing, and compares the sequencing data gatk process and variant comparison.

[0024] Genomic DNA extraction is a routine procedure in molecular biology. Classical methods in molecular biology (e.g., as described in Molecular Cloning) can be used, or various commercially available kits can be used. Samples used for extraction can include muscle, feathers, feces, blood, plasma, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 The distribution of the used SNP sites on the chromosome. DETAILED DESCRIPTION

[0026] Example 1 Baicheng oily chicken variety distribution feature selection and model construction.

[0027] Blood samples were collected from 822 local chicken breeds across China, representing 23 breeds, including 195 Baicheng oily chickens. Compson was commissioned to construct libraries for these blood samples and sequence them on a sequencing platform, generating fastq files for all samples. This study first performed basic quality control on the fastq files using fastp software and quality assessment using fastqc software, ultimately generating clean data. The fastq files were then aligned, deduplicated, sequenced, and variant detected using the GATK pipeline, generating a VCF file containing single-nucleotide polymorphisms (SNPs). This study then performed hard quality control on these VCF files using GATK software, filtering out SNPs of poor quality. The quality control parameters used were: QD < 2.0; QUAL < 30.0; MQ < 40.0; SOR > 3.0; FS > 60.0; MQRankSum < -12.5; and ReadPosRankSum < -8.0.

[0028] After obtaining high-quality VCF files, this study performed soft quality control on the high-quality VCF files. The main quality control steps were: first, only SNP sites on biallelic and autosomal chromosomes were retained, then SNP sites with a missing rate greater than 1% and SNP sites with a minor allele frequency less than 2% were removed to obtain the final VCF file. Next, the mean AED between each population was calculated using the following formula:

[0029]

[0030] Among them, P Aj is the frequency of allele A in the jth variety, and n is the number of varieties. The AED values ​​for each population were calculated, averaged, and sorted from largest to smallest.

[0031] Then, the top 2000 SNPs were extracted from large to small according to the AED value, and the sklearn package in Python was used to perform machine learning feature screening on these SNPs. The model selected was the linear kernel of the support vector machine, C=1.0, and the step size was 5. This method first used all SNP sites to build the model, and then removed the n features (step size 5 in this study) with the least information (SNPs) and built the model again until the 200 features with the highest information content were reached.

[0032] Using this feature selection algorithm, we identified 200 SNPs (shown in Table 1) as characteristic loci for breed assignment. Finally, we divided the training and test sets into a 7:3 ratio, with 575 chickens as the training set and 247 chickens as the test set. The model was built using the training set and tested using the test set, and both the training and test sets were correctly assigned.

[0033] Table 1 Model parameters of support vector machine

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] In the table, the chromosome number represents the chromosome number corresponding to the chicken, the position represents the base number of the SNP on the chromosome, the reference type refers to the base type of the reference genome, the mutant type refers to the base type after gene mutation, the weight refers to the weight of each feature of the machine learning model, and the intercept is the intercept of the machine learning model.

[0047] Example 2 Use of Baicheng Oily Chicken Variety Distribution Model

[0048] Blood was collected from another 10 Baicheng oily chickens, and preliminary work such as library construction and upstream testing was carried out. The obtained VCF file was used to directly extract the 200 SNP sites mentioned above, and the VCF file was converted into a 012 format file using the --recode A parameter in the PLINK software. Missing values ​​were filled with random numbers of 0, 1, and 2, and the SNP information column was extracted to form an SNP feature information file.

[0049] The sklearn package in Python can be used to load the support vector machine model provided in this study or generate a model using the support vector machine parameters provided in this study. The SNP signature information file can then be output to the machine learning model, which will return a prediction result, with 1 indicating a Baicheng Oily Chicken and 0 indicating a non-Baicheng Oily Chicken. In this study, the model was validated using a validation set of 10 Baicheng Oily Chickens from the conservation population and 30 chickens from other breeds. The model successfully identified the 10 Baicheng Oily Chickens from the conservation population as Baicheng Oily Chickens and the other chickens as non-Baicheng Oily Chickens with 100% accuracy, demonstrating the reliability of the machine learning assignment model.

[0050] Table 2 Basic information and test results of machine learning training set, test set and validation set

[0051]

[0052] In actual use, the deletion rate of these sites needs to be below 5%, otherwise it will affect the accuracy of the prediction.

[0053] The above experimental results show that the SNP site combination and machine learning model proposed in this study can well allocate the Baicheng Oily Chicken breeds. At the same time, the model was tested using an exogenous population and obtained good results. This shows that the machine learning model has a good ability to allocate Baicheng Oily Chicken breeds, which provides an important basis for its subsequent application in breed identification.

Claims

1. The SNP combination for identifying the Baicheng oily chicken variety is characterized by: The SNP combination includes the SNP sites shown in Table 1 of the specification.

2. The Baicheng oily chicken variety identification method is characterized by: The method includes using a machine learning model to process the SNP site information shown in Table 1 of the chicken sample to be tested.

3. The method according to claim 2, comprising: (1) Extracting genomic DNA from the chicken sample to be tested; (2) detecting the genotype of the SNP site shown in Table 1 in the genomic DNA obtained in step (1) to obtain the genotype information of the SNP site; (3) Use a machine learning model based on supervised learning to process the SNP site genotype information obtained in step (2).

4. The method according to claim 3, wherein the machine learning model based on supervised learning in step (3) is a linear kernel support vector machine model.

5. The method according to claim 3, wherein in step (2), whole genome resequencing, SNP gene chip method, mass spectrometry, SNaPshot method, KASP method, ARMS method, CAPS method, molecular beacon method, melting curve method, Taqman probe or direct sequencing method are used to detect the genotype of the SNP sites shown in Table 1 in the genomic DNA obtained in step (1).

6. The method according to claim 5, wherein in step (2), whole genome resequencing and SNP gene chip methods are used to detect the genotypes of the SNP sites shown in Table 1 in the genomic DNA obtained in step (1).

7. The method according to claim 3, wherein if there is any missing site information in the SNP site genotype information obtained in step (2), random filling is performed to fill it.

8. Use of a gene chip for detecting the SNP combination according to claim 1 in preparing a Baicheng oily chicken variety identification kit.

9. A method for constructing a Baicheng oily chicken variety identification model based on machine learning, characterized in that: The method comprises: (A) Obtain SNP data for various chicken breeds, including Baicheng Oily Chicken, and perform quality control; (B) Calculate genetic parameters for the quality-controlled SNP data obtained in (A), calculate the Euclidean genetic distance of the quality-controlled SNP data, and rank them accordingly; (C) Some SNP sites with higher European genetic distance were selected to construct the Baicheng oily chicken breed identification model based on machine learning.

10. The method according to claim 9, wherein in (A), SNP data of different duck breeds are obtained by one or more methods selected from database download, whole genome resequencing, and SNP gene chip sequencing; SNP data of different duck breeds are obtained by whole genome resequencing or gene chip sequencing, and the sequencing data are aligned using the gatk process and variant alignment.

Images