基因型特征提取方法影响基因组选择预测准确性的研究

doi:10.11843/j.issn.0366-6964.2024.06.015

摘要/Abstract

摘要：

旨在探索并评估6种不同的单核苷多态性(single nucleotide polymorphisms，SNP)基因型特征提取方法。本研究分析比较了6种方法：主成分分析(principal component analysis，PCA)、基因主成分分析(gene-principal component analysis，gene-PCA)、SNP位点间皮尔逊相关系数(SNP-pearson correlation coefficient, SNP-PCC)、连锁不平衡(linkage disequilibrium，LD)、全基因组关联分析(genome-wide association study，GWAS)和随机抽样(random sampling，RS)，在两组数据(北京鸭，542个样本，SNP位点数39 932；杜洛克猪，2 549个样本，SNP位点数230 884)3组表型(北京鸭体长(body length)、杜洛克猪背膘厚(backfat thickness)和乳头数(teat number))上的GEBV预测准确率。发现SNP-PCC结合5种GS方法(GBLUP、BayesA、BayesB、BayesC、Bayesian Lasso)，在北京鸭数据获得相对可靠的预测精度，在猪背膘厚和乳头数表型获得最高平均预测准确性(提升5%，达到32.3%)，并显著提升计算效率(平均提升5~7倍)。综上，本研究发现选择合适的特征提取方法可以有效提升GS的预测准确性和计算效率，为深入研究不同特征提取方法对GS预测准确性的影响奠定了基础，并为其在育种实践中应用提供了参考。

关键词: 基因组选择, 特征提取, 预测准确性

Abstract:

The purpose of this study was to explore and evaluate 6 different methods for extracting genotype feature of single nucleotide polymorphisms (SNP). Six methods were analyzed and compared: principal component analysis (PCA), gene-principal component analysis (gene-PCA), SNP-Pearson correlation coefficient (SNP-PCC), linkage disequilibrium (LD), and genome-wide association study (GWAS) and random sampling (RS). The prediction accuracy of GEBV in 2 sets of data (Beijing duck, 542 samples, SNP loci 39 932; Duroc pig, 2 549 samples, SNP loci 230 884) and 3 sets of phenotypes (Beijing duck body length, Duroc pig backfat thickness and teat number) was evaluated. Results showed that SNP-PCC combined with 5 GS methods (GBLUP, BayesA, BayesB, BayesC, and Bayesian Lasso) achieved relatively reliable prediction accuracy for the Pecking duck body length phenotype and achieved the highest average prediction accuracy in pig backfat thickness and teat number phenotypes (increased by 5%, reaching 32.3%), and significantly improved computational efficiency (on average 5-7 times faster). In summary, this study found that selecting appropriate feature extraction methods can effectively improve the accuracy and computational efficiency of GS prediction, laying the foundation for in-depth research on the impact of different feature extraction methods on GS prediction accuracy, and providing reference for their application in breeding practice.

Key words: genomic selection, feature extraction, prediction accuracy

中图分类号:

S813.1

吴华煊, 杜志强. 基因型特征提取方法影响基因组选择预测准确性的研究[J]. 畜牧兽医学报, 2024, 55(6): 2431-2440.

Huaxuan WU, Zhiqiang DU. Methods of Genotype Feature Extraction Affecting the Prediction Accuracy of Genomic Selection[J]. Acta Veterinaria et Zootechnica Sinica, 2024, 55(6): 2431-2440.

图/表 7

表 1

图 1

图 2

表 2

图 3

表 3

表 4

参考文献 41

1	MEUWISSENT H E,HAYESB J,GODDARDM E.Prediction of total genetic value using genome-wide dense marker maps[J].Genetics,2001,157(4):1819-1829. doi: 10.1093/genetics/157.4.1819
2	OSTERSENT,CHRISTENSENO F,HENRYONM,et al.Deregressed EBV as the response variable yield more reliable genomic predictions than traditional EBV in pure-bred pigs[J].Genet Sel Evol,2011,43(1):38. doi: 10.1186/1297-9686-43-38
3	ZHAOY S,GOWDAM,LIUW X,et al.Accuracy of genomic selection in European maize elite breeding populations[J].Theor Appl Genet,2012,124(4):769-776. doi: 10.1007/s00122-011-1745-y
4	LIUT F,QUH,LUOC L,et al.Genomic selection for the improvement of antibody response to Newcastle disease and avian influenza virus in chickens[J].PLoS One,2014,9(11):e112685. doi: 10.1371/journal.pone.0112685
5	BEYENEY,SEMAGNK,MUGOS,et al.Genetic gains in grain yield through genomic selection in eight Bi-parental maize populations under drought stress[J].Crop Sci,2015,55(1):154-163. doi: 10.2135/cropsci2014.07.0460
6	PALAIOKOSTASC,FERRARESSOS,FRANCHR,et al.Genomic prediction of resistance to pasteurellosis in gilthead sea bream (Sparus aurata) using 2b-RAD sequencing[J].G3 (Bethesda),2016,6(11):3693-3700. doi: 10.1534/g3.116.035220
7	MEUWISSENT H.Accuracy of breeding values of 'unrelated' individuals predicted by dense SNP genotyping[J].Genet Sel Evol,2009,41(1):35. doi: 10.1186/1297-9686-41-35
8	AKBARZADEHM,DEHKORDIS R,ROUDBARM A,et al.GWAS findings improved genomic prediction accuracy of lipid profile traits: tehran cardiometabolic genetic study[J].Sci Rep,2021,11(1):5780. doi: 10.1038/s41598-021-85203-8
9	LIB,ZHANGN X,WANGY G,et al.Genomic prediction of breeding values using a subset of SNPs identified by three machine learning methods[J].Front Genet,2018,9,237. doi: 10.3389/fgene.2018.00237
10	PILESM,BERGSMAR,GIANOLAD,et al.Feature selection stability and accuracy of prediction models for genomic prediction of residual feed intake in pigs using machine learning[J].Front Genet,2021,12,611506. doi: 10.3389/fgene.2021.611506
11	TORADAL,LORENZONL,BEDDISA,et al.ImaGene: a convolutional neural network to quantify natural selection from genomic data[J].BMC Bioinformatics,2019,20(Suppl 9):337.
12	王万年,陈思佳,郜金荣,等.基于多层感知机的绵羊限性性状基因组选择模拟研究[J].畜牧兽医学报,2023,54(7):2824-2835.
	WANGW N,CHENS J,GAOJ R,et al.Simulation study on genomic selection of sex-limited traits using multilayer perceptron in sheep[J].Acta Veterinaria et Zootechnica Sinica,2023,54(7):2824-2835.
13	丁纪强,李庆贺,张高猛,等.比较机器学习等算法对肉鸡产蛋性状育种值估计的准确性[J].畜牧兽医学报,2022,53(5):1364-1372.
	DINGJ Q,LIQ H,ZHANGG M,et al.Comparing the accuracy of estimated breeding value by several algorithms on laying traits in broilers[J].Acta Veterinaria et Zootechnica Sinica,2022,53(5):1364-1372.
14	AZODIC B,BOLGERE,MCCARRENA,et al.Benchmarking parametric and machine learning models for genomic prediction of complex traits[J].G3 (Bethesda),2019,9(11):3691-3702. doi: 10.1534/g3.119.400498
15	WANGK Q,YANGB,LIQ,et al.Systematic evaluation of genomic prediction algorithms for genomic prediction and breeding of aquatic animals[J].Genes (Basel),2022,13(12):2247. doi: 10.3390/genes13122247
16	XIANGT,LIT,LIJ L,et al.Using machine learning to realize genetic site screening and genomic prediction of productive traits in pigs[J].FASEB J,2023,37(6):e22961. doi: 10.1096/fj.202300245R
17	DENGM T,ZHUF,YANGY Z,et al.Genome-wide association study reveals novel loci associated with body size and carcass yields in Pekin ducks[J].BMC Genomics,2019,20(1):1. doi: 10.1186/s12864-018-5379-1
18	TANC,WUZ F,RENJ L,et al.Genome-wide association study and accuracy of genomic prediction for teat number in Duroc pigs using genotyping-by-sequencing[J].Genet Sel Evol,2017,49(1):35. doi: 10.1186/s12711-017-0311-8
19	GOODFELLOWI,BENGIOY,COURVILLEA.Deep learning[M].Cambridge:The MIT Press,2016.
20	PEDREGOSAF,VAROQUAUXG,GRAMFORTA,et al.Scikit-learn: machine learning in Python[J].J Mach Learn Res,2011,12,2825-2830.
21	SLATKINM.Linkage disequilibrium-understanding the evolutionary past and mapping the medical future[J].Nat Rev Genet,2008,9(6):477-485. doi: 10.1038/nrg2361
22	HILLW G,ROBERTSONA.Linkage disequilibrium in finite populations[J].Theor Appl Genet,1968,38(6):226-231. doi: 10.1007/BF01245622
23	HILLW G,MACKAYT F C.D. S.Falconer and introduction to quantitative genetics[J].Genetics,2004,167(4):1529-1536. doi: 10.1093/genetics/167.4.1529
24	SVEDJ A,HILLW G.One hundred years of linkage disequilibrium[J].Genetics,2018,209(3):629-636.
25	HENDERSONC R.Best linear unbiased estimation and prediction under a selection model[J].Biometrics,1975,31(2):423-447. doi: 10.2307/2529430
26	HABIERD,FERNANDOR L,KIZILKAYAK,et al.Extension of the Bayesian alphabet for genomic selection[J].BMC Bioinformatics,2011,12,186. doi: 10.1186/1471-2105-12-186
27	PÉREZP,DE LOS CAMPOSG.Genome-wide regression and prediction with the BGLR statistical package[J].Genetics,2014,198(2):483-495. doi: 10.1534/genetics.114.164442
28	GREŠOVÁK,MARTINEKV,AČGECHÁK D,et al.Genomic benchmarks: a collection of datasets for genomic sequence classification[J].BMC Genom Data,2023,24(1):1-25. doi: 10.1186/s12863-022-01102-5
29	LUECKENM D,BVTTNERM,CHAICHOOMPUK,et al.Benchmarking atlas-level data integration in single-cell genomics[J].Nature methods,2022,19(1):41-50. doi: 10.1038/s41592-021-01336-8
30	LIY,MANSMANNU,DUS,et al.Benchmark study of feature selection strategies for multi-omics data[J].BMC Bioinformatics,2022,23(1):412. doi: 10.1186/s12859-022-04962-x
31	PRICEA L,PATTERSONN J,PLENGER M,et al.Principal components analysis corrects for stratification in genome-wide association studies[J].Nat Genet,2006,38(8):904-909. doi: 10.1038/ng1847
32	BEHARD M,YUNUSBAYEVB,METSPALUM,et al.The genome-wide structure of the Jewish people[J].Nature,2010,466(7303):238-242. doi: 10.1038/nature09103
33	ATZMONG,HAOL,PE'ERI,et al.Abraham's children in the genome era: major Jewish diaspora populations comprise distinct genetic clusters with shared Middle eastern Ancestry[J].Am J Hum Genet,2010,86(6):850-859. doi: 10.1016/j.ajhg.2010.04.015
34	CAMPBELLC L,PALAMARAP F,DUBROVSKYM,et al.North African Jewish and non-Jewish populations form distinctive, orthogonal clusters[J].Proc Natl Acad Sci U S A,2012,109(34):13865-13870. doi: 10.1073/pnas.1204840109
35	ELHAIKE.Principal component analyses (PCA)-based findings in population genetic studies are highly biased and must be reevaluated[J].Sci Rep,2022,12(1):14683. doi: 10.1038/s41598-022-14395-4
36	REND Y,CAIX D,LINQ,et al.Impact of linkage disequilibrium heterogeneity along the genome on genomic prediction and heritability estimation[J].Genet Sel Evol,2022,54(1):47. doi: 10.1186/s12711-022-00737-3
37	REICHD E,CARGILLM,BOLKS,et al.Linkage disequilibrium in the human genome[J].Nature,2001,411(6834):199-204. doi: 10.1038/35075590
38	CLIMERS,YANGW,DE LAS FUENTESL,et al.A custom correlation coefficient (CCC) approach for fast identification of multi-SNP association patterns in genome-wide SNPs data[J].Genet Epidemiol,2014,38(7):610-621. doi: 10.1002/gepi.21833
39	ZHOUY,VALESM I,WANGA X,et al.Systematic bias of correlation coefficient may explain negative accuracy of genomic prediction[J].Brief Bioinform,2017,18(5):744-753.
40	SUBRAMANIANJ,SIMONR.Overfitting in prediction models-is it a problem only in high dimensions?[J].Contemp Clin Trials,2013,36(2):636-641. doi: 10.1016/j.cct.2013.06.011
41	FROUINA,DANDINE-ROULLANDC,PIERRE-JEANM,et al.Exploring the link between additive heritability and prediction accuracy from a ridge regression perspective[J].Front Genet,2020,11,581594. doi: 10.3389/fgene.2020.581594

表型 Phenotype	样本数 Number of samples	平均值 Mean	标准差 Standard deviation	最大值 Max value	最小值 Min value
体长/cm Body length	542	51.30	2.70	58.00	42.00
背膘厚/mm Backfat thickness	2 549	10.91	2.06	19.70	5.48
乳头数/个Teat number	2 549	10.72	1.06	15.00	8.00

数据集Dataset	SNPs	GWAS(P < 0.05)	PCA	Gene-PCA	LD	SNP-PCC	RS
体长Body length	39 932	4 282	359	542	31 002	26 495	3 993
背膘厚、乳头数 Backfat thickness, teat number	230 884	21 760	716	2 500	33 101	20 731	23 088

数据集 Dataset	方法 Method	平均PCC Mean PCC	PCC标准差 PCC STD	平均MSE Mean MSE	MSE标准差 MSE STD
北京鸭，体长/cm	GWAS(P < 0.05)	0.484	0.081	8.701	2.835
Pecking duck, body length	PCA	0.243	0.043	5.741	0.244
	Gene-PCA	0.132	0.026	6.094	0.080
	LD	0.277	0.052	5.803	0.343
	SNP-PCC	0.290	0.058	5.916	0.556
	Random	0.265	0.036	6.002	0.375
	Origin	0.217	0.020	6.525	0.888
猪，背膘厚/mm	GWAS(P < 0.05)	0.366	0.037	4.927	0.595
Pig, backfat thickness	PCA	0.341	0.034	4.102	0.105
	Gene-PCA	0.186	0.025	4.560	0.118
	LD	0.358	0.020	4.259	0.211
	SNP-PCC	0.367	0.023	4.164	0.181
	Random	0.338	0.021	4.286	0.158
	Origin	0.336	0.026	4.612	0.552
猪，乳头数/个	GWAS(P < 0.05)	0.240	0.042	1.247	0.105
Pig, teat number	PCA	0.313	0.030	1.024	0.029
	Gene-PCA	0.123	0.011	1.148	0.023
	LD	0.301	0.022	1.111	0.068
	SNP-PCC	0.312	0.018	1.090	0.043
	Random	0.299	0.020	1.088	0.053
	Origin	0.280	0.025	1.207	0.158

数据集 Dataset	方法 Method	位点数 Number of markers	计算速度 Computing speed
体长	GWAS(P < 0.05)	4 282	2 s
Body length	PCA	359	1 s
	Gene-PCA	542	1 s
	LD	31 002	10 s
	SNP-PCC	26 495	6 s
	Random	3 993	2 s
	Origin	39 932	15 s
背膘厚、乳头数	GWAS(P < 0.05)	21 760	12 s
Backfat thickness, teat number	PCA	716	1 s
	Gene-PCA	2 500	1.6 s
	LD	33 101	18 s
	SNP-PCC	20 731	11 s
	Random	23 088	12 s
	Origin	230 884	2 min 20 s

[1]	李竟, 张元旭, 王泽昭, 陈燕, 徐凌洋, 张路培, 高雪, 高会江, 李俊雅, 朱波, 郭鹏. 机器学习全基因组选择研究进展[J]. 畜牧兽医学报, 2024, 55(6): 2281-2292.
[2]	钟欣, 张晖, 张充, 刘小红. 母猪繁殖力基因遗传育种研究进展[J]. 畜牧兽医学报, 2024, 55(2): 438-450.
[3]	严晓春, 习海娇, 李金泉, 王志英, 苏蕊. 内蒙古绒山羊绒毛性状基因组育种值估计准确性研究[J]. 畜牧兽医学报, 2024, 55(1): 120-128.
[4]	王万年, 陈思佳, 郜金荣, 温中豪, 袁梦娇, 张洪志, 庞志旭, 乔利英, 刘文忠. 基于多层感知机的绵羊限性性状基因组选择模拟研究[J]. 畜牧兽医学报, 2023, 54(7): 2824-2835.
[5]	李浩东, 闵祥玉, 周雅, 张禾垟, 郑军军, 刘琳玲, 王平, 王艳梅, 杨福合, 王桂武. 基于GBLUP等模型对梅花鹿(Cervus Nippon)生长相关性状基因组选择的预测准确性比较[J]. 畜牧兽医学报, 2023, 54(2): 608-616.
[6]	杨凯, 卢倬达, 何健, 张瑞琪, 王素青, 李克标, 赵云翔, 朱晓萍, 郭金彪. 商品猪群体效应对纯种猪胴体性状基因组选择准确性的影响[J]. 畜牧兽医学报, 2023, 54(12): 4943-4951.
[7]	牛一凡, 杨柏高, 张培培, 张航, 冯肖艺, 曹建华, 余洲, 郝海生, 杜卫华, 邹惠影, 朱化彬, 马友记, 赵学明. 牛胚胎基因组选择研究进展[J]. 畜牧兽医学报, 2023, 54(11): 4449-4457.
[8]	孙东晓, 张胜利, 张勤, 李姣, 张桂香, 刘丑生, 郑伟杰. 我国奶牛基因组选择技术应用进展[J]. 畜牧兽医学报, 2023, 54(10): 4028-4039.
[9]	马浩然, 张路培, 金生云, 宝金山, 李红艳, 高会江, 徐凌洋, 王泽昭, 李俊雅. 利用高密度SNP芯片评估中国地方肉牛品种基因组亲缘关系[J]. 畜牧兽医学报, 2023, 54(10): 4174-4185.
[10]	师睿, 苏国生, 陈紫薇, 李想, 罗汉鹏, 刘林, 郭刚, 张毅, 王雅春, 张胜利, 张勤. 中国荷斯坦牛繁殖性状的基因组预测效果比较[J]. 畜牧兽医学报, 2022, 53(9): 2944-2954.
[11]	庞志旭, 张洪志, 乔利英, 王万年, 潘洋洋, 刘文忠. 基于元共祖的基因组联合育种模拟研究[J]. 畜牧兽医学报, 2022, 53(7): 2172-2181.
[12]	丁纪强, 李庆贺, 张高猛, 李森, 郑麦青, 文杰, 赵桂苹. 比较机器学习等算法对肉鸡产蛋性状育种值估计的准确性[J]. 畜牧兽医学报, 2022, 53(5): 1364-1372.
[13]	杜永旺, 黄超, 王一东, 李森, 文杰, 陈智武, 赵桂苹, 郑麦青. 结合GWAS先验标记信息的肉鸡RFI性状全基因组选择研究[J]. 畜牧兽医学报, 2022, 53(10): 3403-3411.
[14]	张鹏飞, 何俊, 王立贤, 赵福平. 基于基因组和系谱信息的不同选配方案效果模拟研究[J]. 畜牧兽医学报, 2022, 53(10): 3448-3458.
[15]	李森, 杜永旺, 文杰, 黄超, 陈智武, 赵桂苹, 郑麦青. 快速型黄羽肉鸡饲料利用效率性状的基因组选择研究[J]. 畜牧兽医学报, 2021, 52(8): 2151-2161.