基于SNP与机器学习的羊品种识别算法研究

doi:10.3969/j.issn.1004-1524.20241100

浙江农业学报 ›› 2025, Vol. 37 ›› Issue (11): 2387-2394.DOI: 10.3969/j.issn.1004-1524.20241100

基于SNP与机器学习的羊品种识别算法研究

孙硕¹^,²(), 刘昭华³^,⁴, 王可³^,⁴, 郑纪业¹^,²^,^*(), 邢凡彬¹^,², 宋现雪¹^,², 王建英³^,⁴, 孟宪锋³^,⁴, 杨景晁⁵, 张霞¹

1.聊城大学物理科学与信息工程学院,山东聊城 252000
2.山东省农业科学院农业信息与经济研究所,山东济南 250100
3.山东省农业科学院畜牧兽医研究所,山东省畜禽疫病防治与繁育重点实验室,山东济南 250100
4.农业农村部畜禽生物组学重点实验室,山东济南 250100
5.山东省畜牧总站,山东济南 250100

收稿日期:2024-12-20 出版日期:2025-11-25 发布日期:2025-12-08
作者简介:孙硕(2001—),男,山东青岛人,硕士研究生,研究方向为机器学习在农业中的应用。E-mail:stnc4478@126.com
通讯作者: ^*郑纪业,E-mail:jiyezheng@163.com
基金资助:
山东省科技型中小企业创新能力提升工程计划(2024TSGC0082);山东省农业良种工程(2021LGGC010);山东省中央引导地方科技发展资金项目(YDZX2023131)

Breed recognition of sheep based on SNP and machine learning algorithms

SUN Shuo¹^,²(), LIU Zhaohua³^,⁴, WANG Ke³^,⁴, ZHENG Jiye¹^,²^,^*(), XING Fanbin¹^,², SONG Xianxue¹^,², WANG Jianying³^,⁴, MENG Xianfeng³^,⁴, YANG Jingchao⁵, ZHANG Xia¹

1. Department of Physical Science and Information Engineering, Liaocheng University, Liaocheng 252000, Shandong, China
2. Institute of Agricultural Information and Economics, Shandong Academy of Agricultural Sciences, Jinan 250100, China
3. Shandong Key Laboratory of Animal Disease Control and Breeding, Institute of Animal Science and Veterinary Medicine, Shandong Academy of Agricultural Sciences, Jinan 250100, China
4. Key Laboratory of Livestock and Poultry Genomics, Ministry of Agriculture and Rural Affairs, Jinan 250100, China
5. Shandong Provincial Animal Husbandry Station, Jinan 250100, China

Received:2024-12-20 Online:2025-11-25 Published:2025-12-08

摘要/Abstract

摘要： 为确定最优的羊品种识别方法组合,使用11个品种共256头羊的单核苷酸多态性(single nucleotide polymorphism, SNP)基因分型数据,经数据质控后,系统比较3种SNP筛选方法[群体分化指数(fixation index, F_ST)、赋值信息度(informativeness for assignment, In)和最小冗余最大相关(minimum redundancy maximum relevance, mRMR)]和5种机器学习算法[多层感知器(multilayer perceptron, MLP)、极限梯度提升(extreme gradient boosting, XGBoost)、随机森林(random forest, RF)、支持向量机(support vector machine, SVM)和K最邻近法(K-nearest neighbor, KNN)]在不同参考SNP数量条件下的品种识别准确率。结果表明,多数情况下,F_ST筛选效果最佳,SVM算法优势明显,SNP数量对识别准确率影响显著。所有组合中,SVM算法结合F_ST筛选方法,在SNP数量为1 400时效果最佳,品种识别准确率达99.54%。该研究结果有助于理解不同组合下品种识别效果的差异,为保护羊品种多样性和改良特定性状提供帮助。

关键词: 单核苷酸多态性(SNP), 机器学习, 羊, 品种识别, 支持向量机(SVM), 群体分化指数(F_ST)

Abstract:

To explore the optimal combination of methods for sheep breed recognition, we systematically compared the breed recognition accuracy of three single nucleotide polymorphism(SNP) screening methods [fixation index (F_ST), informativeness for assignment (In), and minimum redundancy maximum relevance (mRMR)] and five machine learning algorithms [multilayer perceptron (MLP), extreme gradient boosting (XGBoost), random forest (RF), support vector machine (SVM), and K-nearest neighbor (KNN)] under varying numbers of reference SNPs, using SNPs genotyping data from 256 sheep across 11 breeds after data quality control. The results indicated that in most cases, F_ST demonstrated the best screening performance, the SVM algorithm showed a clear advantage, and the number of SNPs significantly influenced recognition accuracy. Among all combinations, the SVM algorithm combined with the F_ST screening method achieved the best performance with 1 400 SNPs, yielding a breed recognition accuracy of 99.54%. These findings contribute to understanding the differences in breed recognition effectiveness under various combinations and provide support for protecting sheep breed diversity, maintaining ecological balance, and improving specific traits.

Key words: single nucleotide polymorphism(SNP), machine learning algorithm, sheep, breed recognition, support vector machine (SVM), fixation index(F_ST)

中图分类号:

S826.89

孙硕, 刘昭华, 王可, 郑纪业, 邢凡彬, 宋现雪, 王建英, 孟宪锋, 杨景晁, 张霞. 基于SNP与机器学习的羊品种识别算法研究[J]. 浙江农业学报, 2025, 37(11): 2387-2394.

SUN Shuo, LIU Zhaohua, WANG Ke, ZHENG Jiye, XING Fanbin, SONG Xianxue, WANG Jianying, MENG Xianfeng, YANG Jingchao, ZHANG Xia. Breed recognition of sheep based on SNP and machine learning algorithms[J]. Acta Agriculturae Zhejiangensis, 2025, 37(11): 2387-2394.

图/表 5

图1 羊品种识别实验总流程

Fig.1 Overall process for sheep breed recognition testing

图2 总体种群间PCA散点图 Beu,比乌拉羊;BF,獾面羊;BHC,布雷克切维羊;BW,威尔士黑山地羊;HF,威尔士山地羊;HR,拉德诺山地羊;HSF,威尔士斑点羊;Llan,兰韦诺格羊;Lley,莱因羊;LWF,兰多夫白面羊;Tal,塔利邦特威尔士山地羊。下同。

Fig.2 PCA scatter slot of overall populations Beu, Beulah; BF, Badger Face Welsh Mountain; BHC, Brecknock Hill Cheviot; BW, Black Welsh Mountain; HF, Hill Radnor; HR, Hardwick; HSF, Hebridean Speckled Face; Llan, Llanwenog; Lley, Lleyn; LWF, Llandovery Whiteface; Tal, Talybont Welsh Mountain.The same as below.

图3 五种机器学习算法的羊品种识别准确率

Fig.3 Sheep breed recognition accuracy of five machine learning algorithms

图4 不同组合的羊品种识别准确率

Fig.4 Sheep breed recognition accuracy of different combinations

图5 在SVM算法、FST筛选方法和SNP数量为1 400的组合下出现错误预测的混淆矩阵

Fig.5 Confusion matrix of misclassified predictions under the combination of the SVM algorithm, FST selection method and 1 400 SNPs

参考文献 27

[1]	ZHAO C, WANG D, YANG C, et al. Population structure and breed identification of Chinese indigenous sheep breeds using whole genome SNPs and InDels[J]. Genetics Selection Evolution, 2024, 56(1): 60.
[2]	刘丽霞, 张丽, 田晓静, 等. 猪SLA-DRA基因SNP位点筛选及其生物信息学分析[J]. 浙江农业学报, 2017, 29(7): 1077-1085.
	LIU L X, ZHANG L, TIAN X J, et al. SNP screening and bioinformatics analysis of SLA-DRA gene in pigs[J]. Acta Agriculturae Zhejiangensis, 2017, 29(7): 1077-1085. (in Chinese with English abstract)
[3]	LI B, ZHANG N X, WANG Y G, et al. Genomic prediction of breeding values using a subset of SNPs identified by three machine learning methods[J]. Frontiers in Genetics, 2018, 9: 237.
[4]	SHEN Y, HUANG J J, JIA L, et al. Bioinformatics and machine learning driven key genes screening for hepatocellular carcinoma[J]. Biochemistry and Biophysics Reports, 2024, 37: 101587.
[5]	梁卉, 王雪, 司敬方, 等. 利用基因组标记和机器学习算法对中国牛品种的分类准确性研究[J]. 遗传, 2024, 46(7): 530-539.
	LIANG H, WANG X, SI J F, et al. Classification accuracy of machine learning algorithms for Chinese local cattle breeds using genomic markers[J]. Hereditas(Beijing), 2024, 46(7): 530-539. (in Chinese with English abstract)
[6]	BLANCHARD G, BOUSQUET O, ZWALD L. Statistical properties of kernel principal component analysis[J]. Machine Learning, 2007, 66(2): 259-294.
[7]	SCHÖLKOPF B, SMOLA A, MÜLLER K R. Kernel principal component analysis[M]//GERSTNER W, GERMOND A, HASLER M, et al. Artificial neural networks: ICANN'97. Berlin, Heidelberg: Springer Berlin Heidelberg, 1997: 583-588.
[8]	MORADI M H, KHALTABADI-FARAHANI A H, KHODAEI-MOTLAGH M, et al. Genome-wide selection of discriminant SNP markers for breed assignment in indigenous sheep breeds[J]. Annals of Animal Science, 2021, 21(3): 807-831.
[9]	RATHASAMUTH W, PASUPA K, TONGSIMA S. Selection of a minimal number of significant porcine SNPs by an information gain and genetic algorithm hybrid model[EB/OL]. (2019-05-29)[2024-11-30]. https://arxiv.org/abs/1905.09059.
[10]	KASARDA R, MORAVČÍKOVÁ N, MÉSZÁROS G, et al. Classification of cattle breeds based on the random forest approach[J]. Livestock Science, 2023, 267:105143.
[11]	BEYNON S E, SLAVOV G T, FARRÉ M, et al. Population structure and history of the Welsh sheep breeds determined by whole genome genotyping[J]. BMC Genetics, 2015, 16: 65.
[12]	PURCELL S, NEALE B, TODD-BROWN K, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses[J]. American Journal of Human Genetics, 2007, 81(3): 559-575.
[13]	WICKHAM H. Ggplot2[J]. Wiley Interdisciplinary Reviews: Computational Statistics, 2011, 3(2): 180-185.
[14]	ROSENBERG N A, LI L M, WARD R, et al. Informativeness of genetic markers for inference of ancestry[J]. American Journal of Human Genetics, 2003, 73(6): 1402-1422.
[15]	DING C, PENG H C. Minimum redundancy feature selection from microarray gene expression data[J]. Journal of Bioinformatics and Computational Biology, 2005, 3(2): 185-205.
[16]	RIGATTI S J. Random forest[J]. Journal of Insurance Medicine, 2017, 47(1): 31-39.
[17]	DIMITRIADOU E, HORNIK K, LEISCH F, et al. The e1071 package[J]. Misc Functions of Department of Statistics (e1071), TU Wien, 2006: 297-304.
[18]	FARAWAY J J. Extending the linear model with R: generalized linear, mixed effects and nonparametric regression models[M]. 2nd Edition. New York: Chapman and Hall/CRC, 2016.
[19]	LINZER D A, LEWIS J B. poLCA: An R package for polytomous variable latent class analysis[J]. Journal of Statistical Software, 2011, 42(10): 1-29.
[20]	CROOKSTON N L, FINLEY A O. yaImpute: An R package for kNN imputation[J]. Journal of Statistical Software, 2008, 23(10): 1-16.
[21]	CHEN T. XGBoost: extreme gradient boosting[J]. R Package Version 0.4-2, 2015, 1(4).
[22]	RUMELHART D E, HINTON G E, WILLIAMS R J. Learning representations by back-propagating errors[J]. Nature, 1986, 323(6088): 533-536.
[23]	ABADI M, BARHAM P, CHEN J M, et al. TensorFlow: a system for large-scale machine learning[C]//12th USENIX symposium on operating systems design and implementation (OSDI 16), November 2-4, 2016, Savannah, GA. 2016: 265-283.
[24]	SRIVASTAVAN H G, KRIZHEVSKY A. Dropout: a simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15(1): 1929-1958
[25]	SAEYS Y, INZA I, LARRAÑAGA P. A review of feature selection techniques in bioinformatics[J]. Bioinformatics, 2007, 23(19): 2507-2517.
[26]	GILL M, ANDERSON R, HU H F, et al. Machine learning models outperform deep learning models, provide interpretation and facilitate feature selection for soybean trait prediction[J]. BMC Plant Biology, 2022, 22(1): 180.
[27]	XU Z T, DIAO S Q, TENG J Y, et al. Breed identification of meat using machine learning and breed tag SNPs[J]. Food Control, 2021, 125: 107971.

基于SNP与机器学习的羊品种识别算法研究

Breed recognition of sheep based on SNP and machine learning algorithms

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 27

相关文章 15

编辑推荐

Metrics

本文评价

[1]	罗阳兰, 黄丽玲, 黄世旅, 阎勇, 王灿琴. 羊肚菌WSJD-1的基质配方优化及其在广西地区的栽培技术[J]. 浙江农业学报, 2025, 37(9): 1881-1890.
[2]	吴昊霖, 王淑珍, 朱祝军, 何勇. 基于近红外光谱技术和机器学习模型的基质氮含量快速检测[J]. 浙江农业学报, 2025, 37(5): 1159-1171.
[3]	肖银润, 马吉平, 王赟萍, 王素贞, 钟国祥, 熊小文, 张诚. 三种钝化剂对土壤重金属和羊肚菌子实体重金属含量的影响[J]. 浙江农业学报, 2024, 36(7): 1646-1656.
[4]	马丽娜, 田进阳, 王锦, 赵正伟, 马青. 滩羊FGF5、COQ9基因多态性及其与生长性状的关联分析[J]. 浙江农业学报, 2024, 36(5): 1015-1023.
[5]	张永彬, 李想, 满卫东, 刘明月, 樊继好, 胡皓然, 宋利杰, 刘玮佳. 融合Sentinel-1/2数据和机器学习算法的冬小麦产量估算方法研究[J]. 浙江农业学报, 2024, 36(12): 2812-2822.
[6]	左晓洁, 吴明江, 罗琳, 马增岭, 庞观凤, 陈斌斌. 羊栖菜优良品系对高温的耐受性比较[J]. 浙江农业学报, 2024, 36(1): 148-155.
[7]	周丽丽, 冯海宽, 聂臣巍, 许晓斌, 刘媛, 孟麟, 薛贝贝, 明博, 梁齐云, 苏涛, 金秀良. 无人机观测时间对玉米冠层叶绿素密度估算的影响[J]. 浙江农业学报, 2024, 36(1): 18-31.
[8]	杨存明, 刘静, 张梦华, 张晓雪, 刘桂芬, 何军敏, 毛静艺, 李雪, 唐丽, 张文静, 潘林香, 田可川, 黄锡霞. 鲁中肉羊不同生长阶段体尺体重的遗传参数估计[J]. 浙江农业学报, 2024, 36(1): 48-57.
[9]	袁晓春, 王翊帆, 王雅艳, 孙昊然, 孟科, 李新海. 基于RNA测序技术鉴定和分析绵羊毛囊发育相关的可变剪接事件[J]. 浙江农业学报, 2023, 35(9): 2056-2067.
[10]	聂玮, 孟科, 荣轩, 强浩, 郭晨浩, 陶毛孩, 冯登侦. 绵羊GRM1基因多态性及其与肉质性状的相关性[J]. 浙江农业学报, 2023, 35(4): 799-808.
[11]	张梦, 佘宝, 杨玉莹, 黄林生, 朱梦琦. 基于无人机RGB影像的大豆种植区提取方法研究[J]. 浙江农业学报, 2023, 35(4): 952-961.
[12]	温红瑞, 齐明江, 高昌鹏, 田晓雨, 李雪梅, 周玉香. 单宁对育肥滩羊生长性能、消化代谢以及屠宰性能的影响[J]. 浙江农业学报, 2023, 35(12): 2809-2817.
[13]	周力, 桂林生. 饲粮中小麦颗粒用量对藏羊公羔瘤胃内环境的影响[J]. 浙江农业学报, 2023, 35(11): 2543-2554.
[14]	崔玲宇, 喻曼, 乔宇颖, 苏瑶, 王云龙, 沈阿林. 基于Web of Science数据库的土壤质量评价及其微生物指标研究趋势分析[J]. 浙江农业学报, 2023, 35(11): 2688-2697.
[15]	范丽莹, 范婷婷, 仝宗军, 梁立韵, 赵志勇, 陈辉, 周昌艳, 赵晓燕. 镉胁迫对不同品种羊肚菌镉富集规律与抗氧化系统的影响[J]. 浙江农业学报, 2023, 35(10): 2321-2331.