A combination of genetic variation molecular markers for identifying the number of lumbar vertebrae in donkeys and its application.

By screening for combinations of genetic variation molecular markers for the number of lumbar vertebrae in donkeys using machine learning methods, and utilizing 51 genetic variation sites and a random forest model, the problem of identifying the number of lumbar vertebrae in young donkeys in traditional methods has been solved. This approach achieves high-accuracy prediction and low-cost detection for young donkeys, and is applicable to donkey germplasm resource evaluation and breeding.

CN123063531APending Publication Date: 2026-11-10LIAOCHENG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611297469.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-25
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Traditional methods are difficult to accurately identify the number of lumbar vertebrae in young donkeys, and are costly and rely on anatomical observation of adult individuals, which cannot meet the needs of modern molecular breeding.

Method used

Using a combination of genetic variant molecular markers based on machine learning, and leveraging 51 genetic variant sites and a random forest classification model, the number of lumbar vertebrae in donkeys was predicted based on genotype information. This included the detection of 51 genetic variant sites and the application of the random forest model.

Benefits of technology

It achieves high accuracy in predicting the number of lumbar vertebrae in young animals, with an accuracy rate of over 96%. It simplifies the testing process, reduces costs, is suitable for simultaneous testing of large numbers of samples, and improves breeding efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123063531A_ABST
    Figure CN123063531A_ABST
Patent Text Reader

Abstract

This invention relates to a genetic marker combination for identifying the number of lumbar vertebrae in donkeys and its application, belonging to the field of biotechnology. This invention utilizes genome-wide association analysis and machine learning methods to successfully develop a molecular marker combination containing 51 high-contribution genetic variation sites in the Dezhou donkey population. This genetic marker combination, by detecting key mutation sites on specific chromosomes, can efficiently and accurately determine the number of lumbar vertebrae in individual Dezhou donkeys, with an accuracy rate exceeding 96%.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of biotechnology, and in particular to a combination of genetic variation molecular markers for identifying the number of lumbar vertebrae in a donkey and its applications. Background Technology

[0002] The donkey (Equus asinus) is an important traditional livestock resource in my country, with significant economic value in meat, hide, draft, and specialty agricultural industries. With the development of modern animal husbandry, the donkey industry is gradually shifting from traditional breeding methods to large-scale, precision breeding, which places higher demands on the evaluation of donkey germplasm resources, genetic improvement, and selection of superior traits.

[0003] Skeletal structure-related traits are among the important hereditary traits influencing animal body size, production performance, and adaptability. Among these, the composition of the vertebral column, including the number of thoracic and lumbar vertebrae, is a crucial indicator for animal morphological classification, genetic resource evaluation, and evolutionary research. Variation in the number of lumbar vertebrae directly affects the body size and carcass quality of donkeys; individuals with more lumbar vertebrae typically exhibit longer bodies and higher meat yields. Therefore, accurate identification of the lumbar vertebrae phenotype is of significant breeding importance for screening superior germplasm resources and genetic improvement in donkey breeding. The number of lumbar vertebrae (L) in donkeys is typically around 5, but significant individual differences exist within a population. Recent large-scale population surveys of local breeds such as the Dezhou donkey have revealed polymorphic variations in the number of thoracic and lumbar vertebrae. For example, in the Dezhou donkey, the number of thoracic vertebrae (T) ranges from 17 to 19, and the number of lumbar vertebrae ranges from 5 to 6. Traditional animal lumbar vertebrae identification primarily relies on anatomical observation, imaging examination, or morphological measurements of adult individuals. However, current technology still has some shortcomings, such as the detection process relying on adult individuals, which takes a long time; it requires professional equipment and personnel, resulting in high detection costs; the accuracy of identification decreases in some individuals due to age, environmental factors, or observation errors; and it cannot carry out early genetic selection in the early stages, which is not conducive to the construction of modern molecular breeding systems.

[0004] Traditional single-marker association analysis methods primarily evaluate the association between a single genetic variant locus and a target trait through statistical tests. However, for complex quantitative traits regulated by multiple minor genes, the explanatory power of a single locus is limited, and it is easily affected by population structure, sample size, and inter-locus interactions. Therefore, machine learning methods have been increasingly applied to the discovery of genetic markers associated with complex traits in recent years. Machine learning methods can automatically identify genetic features related to the target phenotype from large amounts of genotype data by constructing predictive models. Among them, ensemble learning algorithms such as Random Forest and Gradient Boosting Machine (GBM) construct multiple decision tree models and evaluate the importance of features based on the contribution of different genetic variant loci to the model's classification performance, thereby screening key genetic markers. Compared with traditional association analysis methods, machine learning methods can comprehensively consider the joint contributions of multiple genetic variant loci, making them suitable for the development of molecular markers associated with complex quantitative traits.

[0005] Therefore, machine learning-based genetic variation mining strategies can effectively solve the problem that traditional methods struggle to screen high-contribution markers from massive amounts of genomic variation information, providing a new technical approach for the development of molecular markers related to economic traits in complex animals. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a combination of genetic variation molecular markers for identifying the number of lumbar vertebrae in donkeys and its application.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a combination of genetic variation molecular markers for identifying the number of lumbar vertebrae in a donkey, the combination of genetic variation molecular markers comprising 51 genetic variation sites based on the reference genome ASM1607732v2, the 51 genetic variation sites being as follows: Chromosome 1: locus 6074211; Chromosome 2: Loci 59888947, 62801729, 69966345, 70411243, 105629541, 196206094, 205753276; Chromosome 3: Loci 43117096, 122382956, 142818550; Chromosome 4: locus 39536569; Chromosome 5: loci 27070032, 42521371; Chromosome 6: loci 64195106, 93268335; Chromosome 7: loci 100355706, 114124405, 120245469; Chromosome 8: loci 1833920, 2334164, 20148588; Chromosome 10: locus 13015440; Chromosome 11: locus 23386610; Chromosome 14: locus 21063650; Chromosome 17: loci 24160072, 24476167; Chromosome 20: locus 30584641; Chromosome 21: locus 39768948; Chromosome 22: loci 15053018, 16461042, 37200133; Chromosome 23: locus 6471227; Chromosome 24: loci 6061311, 19720109, 24939458, 33315704; Chromosome 25: locus 14080218; Chromosome 27: locus 3187634; Chromosome 28: loci 145354, 279417, 1262708, 4480135, 16413517, 37602719, 50689919; Chromosome 29: loci 1572327, 30719058, 31161136; Chromosome 30: loci 6924596, 19857204.

[0008] In some specific implementations, the donkey is a Texas donkey.

[0009] Secondly, the present invention provides the application of the aforementioned combination of genetic variation molecular markers in the identification of the number of lumbar vertebrae in donkeys or in the preparation of products for the identification of the number of lumbar vertebrae in donkeys.

[0010] In some specific embodiments, the number of lumbar vertebrae of the donkey includes at least one of five or six lumbar vertebrae.

[0011] Thirdly, the present invention provides a method for predicting the number of lumbar vertebrae in a donkey using the aforementioned combination of genetic variation molecular markers, characterized by comprising the following steps: (1) Collect biological samples from the donkeys to be tested and extract genomic DNA; (2) Detect the genotypes of 51 genetic variation sites in the genomic DNA; (3) Input the genotype information of the 51 genetic variation sites obtained into the pre-established random forest classification model; (4) Based on the output of the random forest classification model, determine whether the individual to be tested belongs to the 5th lumbar vertebra type or the 6th lumbar vertebra type.

[0012] In some specific implementations, the random forest classification model is built using the RandomForest Classifier module in scikit-learn.

[0013] Fourthly, the present invention provides a detection reagent, kit, or detection chip for identifying the number of lumbar vertebrae in a donkey, characterized in that it includes a detection element for detecting 51 genetic variation sites.

[0014] Fifthly, this invention provides the application of a method for combining genetic variation molecular markers or predicting the number of lumbar vertebrae in donkeys in the evaluation of donkey germplasm resources, genetic improvement, or molecular-assisted breeding.

[0015] In a sixth aspect, the present invention provides an electronic device comprising one or more processors and a memory for storing executable instructions, the one or more processors being configured to invoke the executable instructions stored in the memory to implement the steps of a method for predicting the number of lumbar vertebrae in a donkey.

[0016] In a seventh aspect, the present invention provides a computer-readable storage medium storing computer program instructions, which are executed by a processor to implement a method for predicting the number of lumbar vertebrae in a donkey.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention, based on DNA detection, eliminates the need to wait for animals to reach adulthood or undergo anatomical observation, enabling lumbar vertebra number prediction during the juvenile stage. Furthermore, utilizing a combination of 51 genetic variation molecular markers provided by this invention, the accuracy of lumbar vertebra number prediction exceeds 96%, with an AUC of 0.99. In addition, this invention can be combined with microarrays and high-throughput sequencing platforms to achieve simultaneous detection of large numbers of samples, assisting in the screening of Texas donkeys with target skeletal structural characteristics and improving breeding efficiency. Attached Figure Description

[0018] Figure 1 This is a graph showing the accuracy results of the random forest model in identifying the number of lumbar vertebrae in donkeys, as provided in an embodiment of the present invention. Figure 2 The ROC curve of the random forest model provided in the embodiment of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be further described below with reference to specific embodiments. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Various modifications, substitutions, or variations made by those skilled in the art based on the disclosure of the present invention without departing from the technical concept of the present invention are all within the scope of protection of the present invention. Unless otherwise stated, the experimental methods used in the present invention are all conventional experimental methods in the art, and the reagents, instruments, and software used are all commercially available.

[0020] Example 1: Screening and obtaining molecular marker combinations for genetic variations related to the number of lumbar vertebrae in Texas donkeys 1.1 Sample Source: This embodiment selected Texas donkeys with different lumbar vertebra number phenotype records as research subjects. A total of 87 Texas donkey samples were collected. All samples had clear individual numbers, source information, and lumbar vertebra number identification records. Among them, Texas donkeys with different lumbar vertebra number types were included in the analysis (after dissection and counting, the sample population included 49 individuals with 5 lumbar vertebrae and 38 individuals with 6 lumbar vertebrae). Peripheral blood samples were collected from each individual. The samples were used for subsequent analysis after passing DNA quality testing. Animal ethics regulations were followed during the sample collection process.

[0021] 1.2 Genomic reference sequence: This invention uses the donkey reference genome ASM1607732v2 published in the NCBI (National Center for Biotechnology Information) database as the reference genome. All sequencing data were aligned to the above reference genome.

[0022] 1.3 Genomic DNA Extraction: Genomic DNA was extracted using a conventional animal tissue DNA extraction method (Tiangen Biotech (Beijing) Co., Ltd.). The specific steps are as follows: Samples were collected; cell lysis was performed using lysis buffer; proteins were removed by protease digestion; genomic DNA was extracted using magnetic beads; and DNA concentration and integrity were measured. The DNA quality requirements were: A260 / A280 value of 1.8–2.0; DNA concentration meeting sequencing library preparation requirements; and good DNA integrity. The qualified DNA was used for subsequent sequencing analysis.

[0023] 1.4 Resequencing Data Acquisition and Quality Control: Whole-genome resequencing was performed on the BGI-2000 / MGI-T7 sequencing platform to obtain raw sequencing files in FASTQ format. These raw sequencing files were then uploaded to a Linux server for subsequent analysis. FastP software (V0.20.1) was used for quality control of the raw sequencing data. Filtering criteria included: removal of low-quality reads (Q-value <20); removal of adapter-contaminated sequences (reads containing more than 10% N); removal of insufficiently long sequences (reads shorter than 50 bp); and ensuring that the quality of clean reads met the requirements for subsequent alignment. High-quality clean reads were obtained after quality control.

[0024] 1.5 Sequence alignment, variant detection, and preliminary filtering Using BWA software, clean reads were aligned to the ASM1607732v2 donkey reference genome to generate SAM format alignment files. Subsequently, SAMtools was used to convert the SAM files to BAM format, and then sort and indexed them to obtain individual genome coverage and genomic variation information. The HaplotypeCaller module in the GATK (Genome Analysis Toolkit, V4.2.0) software package was used to perform variant detection on individual samples in the sorted BAM files, generating GVCF files. Then, the CombineGVCFs and GenotypeGVCFs modules were used to perform joint genotyping on the GVCF files of all samples to obtain the raw genetic variation set.

[0025] To improve the accuracy of the subsequent lumbar vertebra number prediction model, initial variant sites were filtered. The filtering conditions were as follows: Missing rate filtering, removing sites with genotypic deletions, requiring a missing rate of 0; Allele frequency filtering, using the minimum allele frequency (MAF) ≥ 0.05; Polymorphism information content filtering, using the PIC value ≥ 0.4; Hardy-Weinberg equilibrium test filtering: P < 1 × 10⁻⁶. -5 Abnormal loci; linkage site filtering: for highly linked loci less than 100 kb apart, representative loci were retained to reduce the interference of linkage disequilibrium on subsequent analyses. After the above filtering, a total of 9,432,445 high-quality genetic variation loci were obtained.

[0026] 1.6 Screening of genetic variant sites related to the number of lumbar vertebrae This invention employs machine learning methods for feature selection, including the following steps: Dataset creation: Genotype information of genetic variation sites is used as input variables; the number of lumbar vertebrae in Texas donkeys is used as classification labels to construct a training dataset; Preliminary feature selection: The lumbar vertebra number phenotype (5 or 6 lumbar vertebrae) of 87 Texas donkeys was used as the classification label, with individuals having 5 lumbar vertebrae labeled as 0 and individuals having 6 lumbar vertebrae labeled as 1. Genotypic information of 9,432,445 candidate genetic variation loci was used as the model input variable. The genotypes at each locus were numerically encoded, converting different genotypes into a data matrix recognizable by the model. Homozygous genotypes were encoded as 0, heterozygous genotypes as 1, and homozygous variant alleles as 2. The RandomForestClassifier algorithm from the scikit-learn machine learning toolkit in Python was used for model training. The random forest model parameters were set as follows: n_estimators=5000; oob_score=True; n_jobs=-1; other parameters use default values. Here, `n_estimators` indicates that the number of decision trees in the random forest is set to 5000; `oob_score=True` is used to evaluate the model's generalization ability using out-of-bag samples; and `n_jobs=-1` indicates that all CPU cores are used for computation. Based on the training results of the random forest model, the input sites are ranked according to the feature importance score in the model. Based on the feature importance score output by the model, the top 966 genetic variation sites with the highest scores are selected as candidate core sites, which can be used for subsequent chip design.

[0027] Chromosome-based screening: Since the donkey genome contains multiple chromosomes, to reduce computational complexity, the 966 candidate loci were grouped by chromosome. Gradient Boosting Machine (GBM) models were trained on the genetic variation loci on each chromosome. The specific steps are as follows: The feature importance of candidate loci was evaluated according to chromosomes; loci with high contributions to lumbar vertebra number classification on each chromosome were screened: training was performed iteratively, scoring the contribution of each genetic variation locus in each training iteration, sorting the contributions from largest to smallest, and using the loci with the highest accumulated contributions (reaching 98%) as the training set for the next round of training. This process was repeated iteratively until the training accuracy was no less than 98%; high-contribution loci obtained from different chromosome screenings were merged; the GBM algorithm was used again to evaluate feature contributions; based on model prediction performance and feature contribution values, 51 core genetic variation loci were finally determined as molecular marker combinations for donkey lumbar vertebra number identification. Specific information is shown in Table 1 below. Figure 2 .

[0028] Table 1. Genetic variation markers

[0029]

[0030] Model Selection: To evaluate the stability and predictive ability of the random forest model, the `cross_val_score` function from scikit-learn was used for 10-fold cross-validation. The cross-validation parameters were set as follows: cv=10; scoring='accuracy'; Other parameters use the default settings; The training samples are randomly divided into 10 groups. Each time, 9 groups are selected as the training set, and the remaining group is used as the test set. This process is repeated 10 times, and the model's classification accuracy is calculated. The model evaluation metric is accuracy, calculated using the following formula: Accuracy = Number of correctly predicted samples / Total number of test samples Therefore, using the aforementioned 51 genetic variation sites to establish a random forest prediction model for classifying the number of lumbar vertebrae in 87 training populations, the model's prediction accuracy reached over 96%. This indicates that the selected molecular marker combinations can effectively distinguish between Texas donkeys with 5 and 6 lumbar vertebrae and can be used for molecular identification of the number of lumbar vertebrae in donkeys. (See [link to relevant documentation]). Figure 1 .

[0031] Example 251: Genetic Characterization of Molecular Marker Combinations To further clarify the genetic differences of the 51 loci obtained in Example 1 among Texas donkey populations with different lumbar vertebrae numbers, allele frequency statistical analysis was performed on these 51 loci. The analysis population consisted of 87 Texas donkey individuals obtained in Example 1, including 49 individuals with 5 lumbar vertebrae and 38 individuals with 6 lumbar vertebrae. Based on the genotyping results obtained from whole-genome resequencing, the allele frequency of each locus was calculated in the 5-lumbar vertebrae and 6-lumbar vertebrae populations. The statistical results are shown in Table 2 below.

[0032] Table 2. Statistical results of allele frequencies at loci.

[0033]

[0034]

[0035] The results showed that the 51 loci exhibited different allele frequencies between the 5-lumbar vertebrae and 6-lumbar vertebrae populations. Among them, the dominant alleles of some loci changed significantly in the two populations. The 51 genetic variation loci screened in this invention are genetically associated with the lumbar vertebrae number phenotype in donkeys and can be used as molecular markers for lumbar vertebrae number identification.

[0036] Example 351 Independent Population Validation of Molecular Marker Combinations To verify the accuracy and application effect of the genetic variation molecular marker combination provided by this invention in the actual identification of the number of lumbar vertebrae in Dezhou donkeys, an actual identification verification experiment was carried out using Dezhou donkeys with known lumbar vertebrae phenotype information.

[0037] Experimental materials: A total of 94 Dezhou donkeys were randomly selected from a Dezhou donkey breeding population as validation samples, with no overlap with the aforementioned samples. None of the experimental individuals were used for model training beforehand. Peripheral blood samples were collected from each individual, approximately 2 mL per sample, and stored in anticoagulant tubes. Individual number, sex, age, and origin information were recorded.

[0038] DNA sample collection and sequencing: The blood genomic DNA is extracted using a commercially available animal genomic DNA extraction kit (Thermo Fisher Scientific), which specifically includes: taking 200 μL of whole blood sample; adding lysis solution and protease for cell lysis; purifying DNA through magnetic bead adsorption; washing to remove impurities; eluting to obtain genomic DNA. The DNA quality is detected using a micro-spectrophotometer. A DNA sample is determined to be qualified when the DNA concentration is ≥20 ng / μL, the A260 / A280 value is between 1.8 and 2.0, and electrophoresis shows good DNA integrity. The DNA samples are subjected to resequencing. Based on the sequencing results, the random forest RF model established in the foregoing Example 1 is used to predict the number of lumbar vertebrae. Based on the genotype information of 51 core loci, the prediction result is output for drawing the ROC curve, and the result is compared with the anatomical counting result, as shown in Table 3 below.

[0039] Table 3 Validation results of the independent population

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] The statistical results are shown in Table 4 Table 4 Statistical results

[0047] In summary, a total of 139 samples were detected, and 138 samples were correctly predicted, resulting in a prediction accuracy of 99.28%. Furthermore, based on the predicted probabilities output by the model, ROC curves were plotted with 5 lumbar vertebrae and 6 lumbar vertebrae as positive classes. In the validation population, the AUC for the 5 lumbar vertebrae class was 0.99, and the AUC for the 6 lumbar vertebrae class was also 0.99. This indicates that the combination of 51 genetic variant molecular markers provided by this invention can accurately predict the number and type of lumbar vertebrae in donkeys, and still possesses extremely high classification information in the independent validation population. The model's prediction performance is stable and reliable, demonstrating that this molecular marker combination has high stability and practical application value. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this invention and are not intended to limit the scope of protection of this invention. Although the invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this invention without departing from the essence and scope of the technical solutions of this invention.

Claims

1. A combination of genetic variation molecular markers for identifying the number of lumbar vertebrae in a donkey, characterized in that, The molecular marker combination comprises 51 sites based on the reference genome ASM1607732v2, as shown below: Chromosome 1: locus 6074211; Chromosome 2: Loci 59888947, 62801729, 69966345, 70411243, 105629541, 196206094, 205753276; Chromosome 3: Loci 43117096, 122382956, 142818550; Chromosome 4: locus 39536569; Chromosome 5: loci 27070032, 42521371; Chromosome 6: loci 64195106, 93268335; Chromosome 7: loci 100355706, 114124405, 120245469; Chromosome 8: loci 1833920, 2334164, 20148588; Chromosome 10: locus 13015440; Chromosome 11: locus 23386610; Chromosome 14: locus 21063650; Chromosome 17: loci 24160072, 24476167; Chromosome 20: locus 30584641; Chromosome 21: locus 39768948; Chromosome 22: loci 15053018, 16461042, 37200133; Chromosome 23: locus 6471227; Chromosome 24: loci 6061311, 19720109, 24939458, 33315704; Chromosome 25: locus 14080218; Chromosome 27: locus 3187634; Chromosome 28: loci 145354, 279417, 1262708, 4480135, 16413517, 37602719, 50689919; Chromosome 29: loci 1572327, 30719058, 31161136; Chromosome 30: loci 6924596, 19857204.

2. The molecular marker combination as described in claim 1, characterized in that, The donkey in question is a Texas donkey.

3. The application of the molecular marker combination as described in claim 1 in the identification of the number of lumbar vertebrae in donkeys or in the preparation of products for identifying the number of lumbar vertebrae in donkeys.

4. The application as described in claim 3, characterized in that, The number of lumbar vertebrae in the donkey includes at least one of five or six lumbar vertebrae.

5. A method for predicting the number of lumbar vertebrae in a donkey using the molecular marker combination as described in claim 1, characterized in that, Includes the following steps: (1) Collect biological samples from the donkeys to be tested and extract genomic DNA; (2) Detecting the genotypes of the 51 genetic variation sites described in claim 1 in the genomic DNA; (3) Input the genotype information of the 51 genetic variation sites obtained into the pre-established random forest classification model; (4) Based on the output of the random forest classification model, determine whether the individual to be tested belongs to the 5th lumbar vertebra type or the 6th lumbar vertebra type.

6. The method for predicting the number of lumbar vertebrae in a donkey as described in claim 5, characterized in that, The random forest classification model was built using the Random Forest Classifier module in scikit-learn.

7. A detection reagent, kit, or detection chip for identifying the number of lumbar vertebrae in a donkey, characterized in that, It includes a detection element for detecting the 51 genetic variation sites described in claim 1.

8. The application of the combination of genetic variation molecular markers as described in claim 1 or the method for predicting the number of lumbar vertebrae in donkeys as described in claim 5 in the evaluation of donkey germplasm resources, genetic improvement or molecular-assisted breeding.

9. An electronic device, characterized in that, The electronic device includes one or more processors and a memory for storing executable instructions, the one or more processors being configured to invoke the executable instructions stored in the memory to implement the steps of the method for predicting the number of lumbar vertebrae in a donkey as described in claim 5.

10. A computer-readable storage medium, characterized in that, The storage medium stores computer program instructions, which are executed by a processor to implement the method for predicting the number of lumbar vertebrae in a donkey as described in claim 5.