机器学习全基因组选择研究进展

doi:10.11843/j.issn.0366-6964.2024.06.001

摘要/Abstract

摘要：

机器学习方法是全基因组选择研究的重要分支, 深度学习是近年来机器学习领域新的研究热点。本文介绍了机器学习以及深度学习全基因组选择研究的原理和应用发展, 分别从模型框架、模型参数、特征选择等方面对深度学习全基因组育种值估计研究进展进行了阐述, 探讨了深度学习全基因组选择研究中面临的一些的问题, 并对未来进行了展望。

关键词: 全基因组选择, 研究进展, 机器学习, 深度学习, 原理与应用

Abstract:

Machine learning method is an important branch of genomic selection, and deep learning has become a new research hotspot in the field of machine learning in recent years. The principles and application development of machine learning and deep learning genomic selection were introduced in this paper, and the research progress of deep learning genomic breeding value estimation from the aspects of model framework, model parameters, feature selection were elaborated. Some problems in deep learning genomic selection research were explored and prospects in the future were discussed.

Key words: genomic selection, research progress, machine learning, deep learning, principles and applications

中图分类号:

S813.1

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.

Jing LI, Yuanxu ZHANG, Zezhao WANG, Yan CHEN, Lingyang XU, Lupei ZHANG, Xue GAO, Huijiang GAO, Junya LI, Bo ZHU, Peng GUO. Research Progress in Machine Learning Genomic Selection[J]. Acta Veterinaria et Zootechnica Sinica, 2024, 55(6): 2281-2292.

图/表 4

表 1

表 2

不同卷积神经网络模型性能比较"

模型 Model	数据集 Data set	性状 Trait	试验结果 Experimental results
DeepGS^[19]	Wheat2000	粒长、千粒重等8种性状	准确度指标，粒长性状的DeepGS(0.745)高于FNN(0.378)。平均归一化贴现累积增益值指标，DeepGS在8个性状的结果范围58.98%~445.71%，比传统神经网络高27.70%~246.34；比RR-BLUP高1.44%~65.24%
DNNGP^[47]	Wheat599	四种不同地点下的产量性状	准确度方面，在4种环境下的结果中，DNNGP值最高。在环境1-4中，比GBLUP高64.7%、65.9%、164.2%和61.5%；比LightGBM高36.3%、53.2%、37.2%和38.5%；比SVR高1.4%、14.7%、1.6%和1.5%
	Wheat2000	千粒重、试验重等7种性状	准确度分别比GBLUP、LightGBM、SVR、DeepGS和DLGWAS高234.2%、2.5%、48.9%、16.8%和8.2%
	Maize1401	每穗粒数、每穗粒重、花药日期、PH值	平均准确度而言，SVR最优，DNNGP次优。DNNGP比LightGBM、DeepGS和DNNGP高48.6%、75.1%和167.0%。DNNGP和SVR在花药日期性状的准确性相同，在每穗粒数性状性状中比SVR高12.94%
DualCNN (DLGWAS)^[48]	Soybeans	产量、油脂、高度、水分、蛋白质	DualCNN比DeepGS、singleCNN、rrBLUP、BayesA、BL、BRR的准确度分别高2%、2.4%、2.4%、0.7%和1%
ResGS^[49]	Wheat599	四种不同地点下的产量	就准确度而言，在1，2，4环境中比FNN、DeepGS和GBLUP高4.06%~101.59%、2.24%~130.83%和1.71%~107.21%；在环境3中比FNN和DeepGS高20.47%和1.76%，比GBLUP低1.3%
ResGS^[50]	Rice413	蛋白质含量、种子长度等6种性状	平均准确度而言，ResGS结果最高(0.75)；DNNGP(0.71)次之；RRBLUP、SVR、RF和GBR的结果范围0.61~0.65
	Rice395	直链淀粉含量、种子长度	直链淀粉含量性状，ResGS(0.94)、DNNGP(0.83)、RF(0.89)和GBR(0.88)；种子长度性状，ResGS(0.84)、DNNGP(0.85)、RF(0.89)和GBR(0.88)
	Maize301	授粉天数、穗直径、穗高	平均预测准确度而言，ResGS和DNNGP比RRBLUP、SVR、RF、GBR高10%以上；单性状而言，ResGS的准确度分别为0.78、0.65和0.56；DNNGP分别为0.78、0.62和0.57
SoyDNGP^[51]	Soybeans	株高、含油量等7种性状	平均预测准确度而言，DNNGP与SoyDNGP相差约5%；均方误差而言，SoyDNGP比DNNGP高10%
	Cotton1039	铃重等5种性状	SoyDNGP的准确度范围约为0.50~0.70；DNNGP约为0.49~0.69在Mazize、Tomato中DNNGP略高于SoyDNGP；在Cotton、Rice中SoyDNGP略高于DNNGP
	Rice1765	茎长等5种性状
	Mazize508	穗高等5种性状
	Tomato214	茎长、抽穗天数等5种性状

表 2

表 3

表 4

参考文献 98

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	WELLERJ I,EZRAE,RONM.Invited review: A perspective on the future of genomic selection in dairy cattle[J].J Dairy Sci,2017,100(11):8633-8644. doi: 10.3168/jds.2017-12879
3	VARGASR,MOSAVIA,RUIZR.Deep learning: a review[J].Adv Intell Syst Comput,2017,5(2)
4	ZÖLLERM A,HUBERM F.Benchmark and survey of automated machine learning frameworks[J].J Artif Intell Res,2021,70,409-472. doi: 10.1613/jair.1.11854
5	ALVESA A C,ESPIGOLANR,BRESOLINT,et al.Genome-enabled prediction of reproductive traits in Nellore cattle using parametric models and machine learning methods[J].Anim Genet,2021,52(1):32-46. doi: 10.1111/age.13021
6	KARACAÖRENB.An evaluation of machine learning for genomic prediction of hairy syndrome in dairy cattle[J].Anim Sci Pap Rep,2022,40(1):45-58.
7	WANGX,SHIS L,WANGG J,et al.Using machine learning to improve the accuracy of genomic prediction of reproduction traits in pigs[J].J Anim Sci Biotechnol,2022,13(1):60. doi: 10.1186/s40104-022-00708-0
8	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.
9	ZHOU Z H. Ensemble learning[M]//ZHOU Z H. Machine Learning. Singapore: Springer, 2021: 181-210.
10	FREUNDY,SCHAPIRER E.A short introduction to boosting[J].J JSAI,1999,14(5):771-780.
11	GONZÁLEZ-RECIOO,FORNIS.Genome-wide prediction of discrete traits using Bayesian regressions and machine learning[J].Genet Sel Evol,2011,43(1):7. doi: 10.1186/1297-9686-43-7
12	ABDOLLAHI-ARPANAHIR,GIANOLAD,PEÑAGARICANOF.Deep learning versus parametric and ensemble methods for genomic prediction of complex phenotypes[J].Genet Sel Evol,2020,52(1):12. doi: 10.1186/s12711-020-00531-z
13	GRINBERGN F,ORHOBORO I,KINGR D.An evaluation of machine-learning for predicting phenotype: studies in yeast, rice, and wheat[J].Mach Learn,2020,109(2):251-277. doi: 10.1007/s10994-019-05848-5
14	BREIMANL.Bagging predictors[J].Mach Learn,1996,24(2):123-140.
15	JAMESG,WITTEND,HASTIET,et al.An introduction to statistical learning: with Applications in R[M].New York:Springer,2013.
16	SILVEIRAL S,LIMAL P,NASCIMENTOM,et al.Regression trees in genomic selection for carcass traits in pigs[J].Genet Mol Res,2020,19(1):gmr18498.
17	GIANOLAD,WEIGELK A,KRÄMERN,et al.Enhancing genome-enabled prediction by bagging genomic BLUP[J].PLoS One,2014,9(4):e91693. doi: 10.1371/journal.pone.0091693
18	梁忙. 基于机器学习算法的全基因组选择研究[D]. 北京: 中国农业科学院, 2021.
	LIANG M. The algorithm research for genomic selection study based on machine learning[D]. Beijing: Chinese Academy of Agricultural Sciences, 2021. (in Chinese)
19	MAW L,QIUZ X,SONGJ,et al.DeepGS: Predicting phenotypes from genotypes using deep learning[J].BioRxiv,2017,241414.
20	ZHAOW,LAIX S,LIUD Y,et al.Applications of support vector machine in genomic prediction in pig and maize populations[J].Front Genet,2020,11,598318. doi: 10.3389/fgene.2020.598318
21	DE LOS CAMPOSG,GIANOLAD,ROSAG J M.Reproducing kernel Hilbert spaces regression: a general framework for genetic evaluation[J].J Anim Sci,2009,87(6):1883-1887. doi: 10.2527/jas.2008-1259
22	BLONDELM,ONOGIA,IWATAH,et al.A ranking approach to genomic selection[J].PLoS One,2015,10(6):e0128570. doi: 10.1371/journal.pone.0128570
23	GONZÁLEZ-RECIOO,ROSAG J M,GIANOLAD.Machine learning methods and predictive ability metrics for genome-wide prediction of complex traits[J].Livest Sci,2014,166,217-231. doi: 10.1016/j.livsci.2014.05.036
24	GONZÁLEZ-CAMACHOJ M,ORNELLAL,PÉREZ-RODRÍGUEZP,et al.Applications of machine learning methods to genomic selection in breeding wheat for rust resistance[J].Plant Genome,2018,11(2):170104. doi: 10.3835/plantgenome2017.11.0104
25	XUY,XUC,XUS.Prediction and association mapping of agronomic traits in maize using multiple omic data[J].Heredity (Edinb),2017,119(3):174-184. doi: 10.1038/hdy.2017.27
26	GONZÁLEZ-RECIOO,GIANOLAD,LONGN Y,et al.Nonparametric methods for incorporating genomic information into genetic evaluations: an application to mortality in broilers[J].Genetics,2008,178(4):2305-2313. doi: 10.1534/genetics.107.084293
27	ZOUH,HASTIET.Regularization and variable selection via the elastic net[J].J R Stat Soc Ser B,2005,67(2):301-320. doi: 10.1111/j.1467-9868.2005.00503.x
28	WANGX Q,MIAOJ,CHANGT P,et al.Evaluation of GBLUP, BayesB and elastic net for genomic prediction in Chinese Simmental beef cattle[J].PLoS One,2019,14(2):e0210442. doi: 10.1371/journal.pone.0210442
29	ANB X,LIANGM,CHANGT P,et al.K_CRR: a nonlinear machine learning with a modified genomic similarity matrix improved the genomic prediction efficiency[J].Brief Bioinform,2021,22(6):bbab132. doi: 10.1093/bib/bbab132
30	GENTZBITTELL,BENC,MAZURIERM,et al.WhoGEM: An admixture-based prediction machine accurately predicts quantitative functional traits in plants[J].Genome Biol,2019,20(1):106. doi: 10.1186/s13059-019-1697-0
31	YINL L,ZHANGH H,ZHOUX,et al.KAML: improving genomic prediction accuracy of complex traits using machine learning determined parameters[J].Genome Biol,2020,21(1):146. doi: 10.1186/s13059-020-02052-w
32	TODAY,WAKATSUKIH,AOIKET,et al.Predicting biomass of rice with intermediate traits: Modeling method combining crop growth models and genomic prediction models[J].PLoS One,2020,15(6):e0233951. doi: 10.1371/journal.pone.0233951
33	VAN DER HEIDEE M M,VEERKAMPR F,VAN PELTM L,et al.Comparing regression, naive Bayes, and random forest methods in the prediction of individual survival to second lactation in Holstein cattle[J].J Dairy Sci,2019,102(10):9409-9421. doi: 10.3168/jds.2019-16295
34	GILLBERGJ,MARTTINENP,MAMITSUKAH,et al.Modelling G×E with historical weather information improves genomic prediction in new environments[J].Bioinformatics,2019,35(20):4045-4052. doi: 10.1093/bioinformatics/btz197
35	BELLOTP,DE LOS CAMPOSG,PÉREZ-ENCISOM.Can deep learning improve genomic prediction of complex human traits?[J].Genetics,2018,210(3):809-819. doi: 10.1534/genetics.118.301298
36	XIE B, ZHANG Q. Deep filtering with DNN, CNN and RNN[J]. arXiv preprint arXiv: 2112.12616v2, 2009.
37	束永俊,吴磊,王丹,等.人工神经网络在作物基因组选择中的应用[J].作物学报,2011,37(12):2179-2186.
	SHUY J,WUL,WANGD,et al.Application of artificial neural network in genomic selection for crop improvement[J].Acta Agronomica Sinica,2011,37(12):2179-2186.
38	MONTESINOS-LÓPEZA,MONTESINOS-LÓPEZO A,GIANOLAD,et al.Multi-environment genomic prediction of plant traits using deep learners with dense architecture[J].G3 (Bethesda),2018,8(12):3813-3828. doi: 10.1534/g3.118.200740
39	KHAKIS,WANGL Z.Crop yield prediction using deep neural networks[J].Front Plant Sci,2019,10,621. doi: 10.3389/fpls.2019.00621
40	RANZATO M A, BOUREAU Y L, LECUN Y. Sparse feature learning for deep belief networks[C]//Proceedings of the 20th International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc., 2007: 1185-1192.
41	ISLAMT,KIMC H,IWATAH,et al.DeepCGP: a deep learning method to compress genome-wide polymorphisms for predicting phenotype of rice[J].IEEE/ACM Trans Comput Biol Bioinform,2023,20(3):2078-2088. doi: 10.1109/TCBB.2022.3231466
42	LOPES N, RIBEIRO B. Deep belief networks (DBNs)[M]//LOPES N, RIBEIRO B. Machine Learning for Adaptive Many-Core Machines-A Practical Approach. Cham: Springer, 2015: 155-186.
43	RACHMATIAH,KUSUMAW A,HASIBUANL S.Prediction of maize phenotype based on whole-genome single nucleotide polymorphisms using deep belief networks[J].J Phys: Conf Ser,2017,835,012003. doi: 10.1088/1742-6596/835/1/012003
44	YAMASHITAR,NISHIOM,DOR K G,et al.Convolutional neural networks: an overview and application in radiology[J].Insights Imaging,2018,9(4):611-629. doi: 10.1007/s13244-018-0639-9
45	GHOLAMALINEZHAD H, KHOSRAVI H. Pooling methods in deep neural networks, a review[J]. arXiv preprint arXiv: 2009.07481, 2020.
46	MONTGOMERYD C,PECKE A,VININGG G.Introduction to linear regression analysis[M].6th ed.Hoboken:John Wiley & Sons,2021.
47	WANGK L,ABIDM A,RASHEEDA,et al.DNNGP, a deep neural network-based method for genomic prediction using multi-omics data in plants[J].Mol Plant,2023,16(1):279-293. doi: 10.1016/j.molp.2022.11.004
48	LIUY,WANGD L,HEF,et al.Phenotype prediction and genome-wide association study using deep convolutional neural network of soybean[J].Front Genet,2019,10,1091. doi: 10.3389/fgene.2019.01091
49	顾林林. 基于人工智能(AI)技术的基因组遗传值预测的新算法开发[D]. 厦门: 集美大学, 2021.
	GU L L. Development of new algorithms for genetic value prediction of genomes based on artificial intelligence (AI) techniques[D]. Xiamen: Jimei University, 2021. (in Chinese)
50	XIEZ C,XUX G,LIL,et al.Residual networks without pooling layers improve the accuracy of genomic predictions[J].Theor Appl Genet,2023,
51	GAOP F,ZHAOH N,LUOZ,et al.SoyDNGP: a web-accessible deep learning framework for genomic prediction in soybean breeding[J].Brief Bioinform,2023,24(6):bbad349. doi: 10.1093/bib/bbad349
52	YANGL,SHAMIA.On hyperparameter optimization of machine learning algorithms: Theory and practice[J].Neurocomputing,2020,415,295-316. doi: 10.1016/j.neucom.2020.07.061
53	LIANGM,ANB X,LIK,et al.Improving genomic prediction with machine learning incorporating TPE for hyperparameters optimization[J].Biology (Basel),2022,11(11):1647.
54	HANJ J,GONDROC,REIDK,et al.Heuristic hyperparameter optimization of deep learning models for genomic prediction[J].G3 (Bethesda),2021,11(7):jkab032. doi: 10.1093/g3journal/jkab032
55	STORNR,PRICEK.Differential evolution-a simple and efficient heuristic for global optimization over continuous spaces[J].J Global Optim,1997,11(4):341-359. doi: 10.1023/A:1008202821328
56	AlVESA A C,FERNANDESA F A,LOPESF B,et al.(Quasi) multitask support vector regression with heuristic hyperparameter optimization for whole-genome prediction of complex traits: a case study with carcass traits in broilers[J].G3 (Bethesda),2023,13(8):jkad109. doi: 10.1093/g3journal/jkad109
57	MIRJALILI S. Genetic algorithm[M]//MIRJALILI S. Evolutionary Algorithms and Neural Networks: Theory and Applications. Cham: Springer, 2019: 43-55.
58	WALDMANNP,PFEIFFERC,MÉSZÁROSG.Sparse convolutional neural networks for genome-wide prediction[J].Front Genet,2020,11,25. doi: 10.3389/fgene.2020.00025
59	FALKNER S, KLEIN A, HUTTER F. BOHB: Robust and efficient hyperparameter optimization at scale[C]//Proceedings of the 35th International Conference on Machine Learning. Stockholm: PMLR, 2018: 1436-1445.
60	ZHOUG H,GAOJ,ZUOD S,et al.MSXFGP: combining improved sparrow search algorithm with XGBoost for enhanced genomic prediction[J].BMC Bioinformatics,2023,24(1):384. doi: 10.1186/s12859-023-05514-7
61	XUEJ K,SHENB.A novel swarm intelligence optimization approach: sparrow search algorithm[J].Syst Sci Control Eng,2020,8(1):22-34. doi: 10.1080/21642583.2019.1708830
62	OKUTH,WUX L,ROSAG J M,et al.Predicting expected progeny difference for marbling score in Angus cattle using artificial neural networks and Bayesian regression models[J].Genet Sel Evol,2013,45(1):34. doi: 10.1186/1297-9686-45-34
63	WALDMANNP.Approximate Bayesian neural networks in genomic prediction[J].Genet Sel Evol,2018,50(1):70. doi: 10.1186/s12711-018-0439-1
64	GONZÁLEZ-CAMACHOJ M,DE LOS CAMPOSG,PÉREZP,et al.Genome-enabled prediction of genetic values using radial basis function neural networks[J].Theor Appl Genet,2012,125(4):759-771. doi: 10.1007/s00122-012-1868-9
65	GONZÁLEZ-CAMACHOJ M,CROSSAJ,PÉREZ-RODRÍGUEZP,et al.Genome-enabled prediction using probabilistic neural network classifiers[J].BMC Genomics,2016,17,208. doi: 10.1186/s12864-016-2553-1
66	JUBAIR S, DOMARATZKI M. Ensemble supervised learning for genomic selection[C]//2019 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). San Diego: IEEE, 2019: 1993-2000.
67	CROSSAJ,JARQUÍND,FRANCOJ,et al.Genomic prediction of gene bank wheat landraces[J].G3 (Bethesda),2016,6(7):1819-1834. doi: 10.1534/g3.116.029637
68	DONGS,WANGP,ABBASK.A survey on deep learning and its applications[J].Comput Sci Rev,2021,40,100379. doi: 10.1016/j.cosrev.2021.100379
69	LECUNY,BENGIOY,HINTONG.Deep learning[J].Nature,2015,521(7553):436-444. doi: 10.1038/nature14539
70	JUBAIRS,TUCKERJ R,HENDERSONN,et al.GPTransformer: A transformer-based deep learning method for predicting Fusarium related traits in barley[J].Front Plant Sci,2021,12,761402. doi: 10.3389/fpls.2021.761402
71	WASHBURNJ D,CIMENE,RAMSTEING,et al.Predicting phenotypes from genetic, environment, management, and historical data using CNNs[J].Theor Appl Genet,2021,134(12):3997-4011. doi: 10.1007/s00122-021-03943-7
72	JUBAIRS,DOMARATZKIM.Crop genomic selection with deep learning and environmental data: A survey[J].Front Artif Intell,2023,5,1040295. doi: 10.3389/frai.2022.1040295
73	ALWOSHEELA,VAN CRANENBURGHS,CHORUSC G.Is your dataset big enough? Sample size requirements when using artificial neural networks for discrete choice analysis[J].J Choice Model,2018,28,167-182. doi: 10.1016/j.jocm.2018.07.002
74	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: transformers for image recognition at scale[C]//9th International Conference on Learning Representations. ICLR, 2021.
75	AZIZI S, MUSTAFA B, RYAN F, et al. Big self-supervised models advance medical image classification[C]//Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Montreal: IEEE, 2021: 3458-3468.
76	WANGJ L.Fast and accurate population admixture inference from genotype data from a few microsatellites to millions of SNPs[J].Heredity,2022,129(2):79-92. doi: 10.1038/s41437-022-00535-z
77	NAYERIS,SARGOLZAEIM,TULPAND.A review of traditional and machine learning methods applied to animal breeding[J].Anim Health Res Rev,2019,20(1):31-46. doi: 10.1017/S1466252319000148
78	MONTESINOS-LÓPEZO A,CRESPO-HERRERAL,PIERREC S,et al.Do feature selection methods for selecting environmental covariables enhance genomic prediction accuracy?[J].Front Genet,2023,14,1209275. doi: 10.3389/fgene.2023.1209275
79	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
80	PUDJIHARTONON,FADASONT,KEMPA-LIEHRA W,et al.A review of feature selection methods for machine learning-based disease risk prediction[J].Front Bioinform,2022,2,927312. doi: 10.3389/fbinf.2022.927312
81	LIS S,YUJ,KANGH M,et al.Genomic selection in Chinese Holsteins using regularized regression models for feature selection of whole genome sequencing data[J].Animals (Basel),2022,12(18):2419.
82	ISLAM T, KIM C H, IWATA H, et al. A deep learning method to impute missing values and compress genome-ide polymorphism data in rice[C]//Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies. BIOSTEC, 2021: 101-109.
83	ERASLANG,AVSECŽ,GAGNEURJ,et al.Deep learning: new computational modelling techniques for genomics[J].Nat Rev Genet,2019,20(7):389-403.
84	DE CONINCKA,FOSTIERJ,MAENHOUTS,et al.DAIRRy-BLUP: A high-performance computing approach to genomic prediction[J].Genetics,2014,197(3):813-822. doi: 10.1534/genetics.114.163683
85	GUOP,ZHUB,NIUH,et al.Fast genomic prediction of breeding values using parallel Markov chain Monte Carlo with convergence diagnosis[J].BMC Bioinformatics,2018,19(1):3. doi: 10.1186/s12859-017-2003-3
86	CHENL H,LIC X,MILLERS,et al.Multi-population genomic prediction using a multi-task Bayesian learning model[J].BMC Genetics,2014,15,53.
87	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 770-778.
88	SRIVASTAVAN,HINTONG,KRIZHEVSKYA,et al.Dropout: a simple way to prevent neural networks from overfitting[J].J Mach Learn Res,2014,15(1):1929-1958.
89	IOFFE S, SZEGEDY C. Batch normalization: Accelerating deep network training by reducing internal covariate shift[C]//Proceedings of the 32nd International Conference on International Conference on Machine Learning. Lille: JMLR. org., 2015: 448-456.
90	SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//3rd International Conference on Learning Representations. San Diego: ICLR, 2015.
91	KIZILKAYAK,FERNANDOR L,GARRICKD J.Genomic prediction of simulated multibreed and purebred performance using observed fifty thousand single nucleotide polymorphism genotypes[J].J Anim Sci,2010,88(2):544-551. doi: 10.2527/jas.2009-2064
92	ERBEM,HAYESB J,MATUKUMALLIL K,et al.Improving accuracy of genomic predictions within and between dairy cattle breeds with imputed high-density single nucleotide polymorphism panels[J].J Dairy Sci,2012,95(7):4114-4129. doi: 10.3168/jds.2011-5019
93	LEEH J,LEEJ H,GONDROC,et al.deepGBLUP: joint deep learning networks and GBLUP framework for accurate genomic prediction of complex traits in Korean native cattle[J].Genet Sel Evol,2023,55(1):56. doi: 10.1186/s12711-023-00825-y
94	POOKT,FREUDENTHALJ,KORTEA,et al.Using local convolutional neural networks for genomic prediction[J].Front Genet,2020,11,561497. doi: 10.3389/fgene.2020.561497
95	LUJ,HOUW,XIONGL W,et al.GSCNN: A genomic selection convolutional neural network model based on SNP genotype and physical distance features and data augmentation strategy[J].BMC Genomics,2024,
96	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 2017: 6000-6010.
97	WUZ H,PANS R,CHENF W,et al.A comprehensive survey on graph neural networks[J].IEEE Trans Neural Netw Learn Syst,2021,32(1):4-24. doi: 10.1109/TNNLS.2020.2978386
98	HAMILTON W L, YING Z, LESKOVEC J. Inductive representation learning on large graphs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 2017: 1025-1035.

机器学习模型 Machine Learning method	技术特点 Technical feature	物种 Species
EN^[28]	综合了RR和LASSO两种模型，通过参数调节RR和LASSO在EN中影响比重	华西牛
KcRR^[29]	利用余弦核函数替换岭回归核函数	华西牛、火炬松
KNN^[6]	使用欧几里得距离表示个体间SNP的距离，选取最近邻的K个个体估计育种值	奶牛
WhoGEM^[30]	在GS中加入位置信息作为协变量帮助获取混合物成分的最优值用于GS的预测	截形苜蓿
KAML^[31]	通过整合伪QTN作为协变量和优化的性状特异性随机效应扩展LMM；选择具有显著效应的SNP作为协变量构建亲缘关系矩阵，根据效应大小为SNP分配不同权重; 所有未知参数通过交叉验证、多元回归、网格搜索和二分算法等进行优化	牛、马、玉米
DVR^[32]	利用GP预测相关性状基因组育种值，结合DVR模型、基因组数据、环境数据估计目标性状基因组育种值	日本粳稻
NB^[33]	利用先验概率形构成简单贝叶斯网络; 利用简单贝叶斯网络估计育种值	荷斯坦牛
KBMF^[34]	利用KBMF(核化贝叶斯分解)预测未来生长季节天气条件，实现目标性状的基因型-环境互作效应的育种值估计	大麦

模型 Model	KRR^[53] SVR^[53]	MLP^[54] CNN^[54]	QMTSVR^[56]	CNNGWP^[58]	MSXFGP^[60]
优化方法 Hyper Parameter Optimization	树结构-贝叶斯优化 TPE^[52]	差分进化 DE^[55]	遗传算法 GA^[57]	贝叶斯优化 BO^[59]	麻雀算法 SSA^[61]
优化的超参数 Optimized Hyperparameter	SVR和KRR核函数 Gamma值 Alpha值 K值	激活函数隐层数神经元个数批次 Epoch值 Dropout值 L2值	ρ值核宽带 C值	卷积核个数核大小 L1值	学习率树最大深度子节点最小权重样本子样本比例列样本子样本比例

神经网络 Neural networks	正则化 Regularized	物种 Species	性状 Trait	试验结果 Experimental results
BRANN^[62]	贝叶斯	安格斯牛	大理石花纹评分	BRANN的SSE约为传统的 SCGANN的40%至50%
ABNN^[63]	Dropout、L1、贝叶斯	猪	t3 (h²=0.38)	不同权重衰竭下ABNN的MSE范围0.8653~0.8688、GBLUP(0.8759)、BLASSO(0.8741)
RBFNN^[64]	BRF函数	玉米	雌雄性开花、粮食产量等21种性状	平均准确性而言，RNFNN(0.547)、RKHS(0.553)、BL(0.542)
PNN^[65]	竞争函数	玉米	在高产环境下、水源充足下的产量等16种性状	AUC而言，上层和下层(15%和30%)和中层(40%和70%)类别中选择的性状，PNN结果优于浅层MLP
PNN^[65]	竞争函数	小麦	在干旱下的产量、全灌床下的抽穗期天数等17种性状	AUC而言，上层和下层(15%和30%)和中层(40%和70%)类别中选择的性状，PNN结果优于浅层MLP

[1]	张元旭, 李竟, 王泽昭, 陈燕, 徐凌洋, 张路培, 高雪, 高会江, 李俊雅, 朱波, 郭鹏. 动物遗传评估软件研究进展[J]. 畜牧兽医学报, 2024, 55(5): 1827-1841.
[2]	彭佩雅, 陈钰焓, 杨龙, 王铭, 赵芮葶, 何俊, 印遇龙, 刘梅. 家畜基因组拷贝数变异研究进展[J]. 畜牧兽医学报, 2024, 55(4): 1356-1369.
[3]	周梦婷, 宋银娟, 许健, 李斌, 冉多良, 储岳峰. 基于碳水化合物的佐剂作用机制研究进展[J]. 畜牧兽医学报, 2024, 55(2): 491-501.
[4]	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.
[5]	褚楚, 张静静, 丁磊, 樊懿楷, 包向男, 向世馨, 刘锐, 罗雪路, 任小丽, 李春芳, 刘文举, 王亮, 刘莉, 李永青, 江汉, 李委奇, 孙伟, 李喜和, 温万, 周佳敏, 张淑君. 基于中红外光谱的牛奶中三种氨基酸含量预测模型的建立及应用[J]. 畜牧兽医学报, 2023, 54(8): 3299-3312.
[6]	夏春秋, 万发春, 刘磊, 沈维军, 肖定福. 缬氨酸的生物学功能及其在畜禽日粮中的应用[J]. 畜牧兽医学报, 2023, 54(11): 4502-4513.
[7]	毕睿晨, 刘相泽, 胡泽琼, 杨美雪, 乔嘉宁, 黄嘉, 郭芳申, 孔令华, 王忠. 植物多酚在家禽领域应用研究进展[J]. 畜牧兽医学报, 2023, 54(11): 4488-4501.
[8]	孙康泰, 刘彬, 刘军, 蒋大伟, 姚志鹏, 葛毅强, 邓小明. 基于文献计量的“十三五”国家重点研发计划“畜禽专项”研究进展与热点趋势分析[J]. 畜牧兽医学报, 2022, 53(9): 2819-2832.
[9]	赵志显, 常雪蕊, 郭勇, 龙城, 盛熙晖, 王相国, 邢凯, 肖龙菲, 林自力, 倪和民, 齐晓龙. 种公鸡精液品质营养调控的研究进展[J]. 畜牧兽医学报, 2022, 53(8): 2435-2443.
[10]	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.
[11]	朱波, 李宏伟, 周姵诺, 李倩, 高翰, 王泽昭, 汪聪勇, 蔡文涛, 徐凌洋, 陈燕, 张路培, 高雪, 高会江, 李俊雅. 国内外肉牛遗传评估体系概况[J]. 畜牧兽医学报, 2021, 52(6): 1447-1460.
[12]	袁泽湖, 葛玲, 李发弟, 乐祥鹏, 孙伟. 整合生物学先验信息的全基因组选择方法及其在家畜育种中的应用进展[J]. 畜牧兽医学报, 2021, 52(12): 3323-3334.
[13]	尹立林, 马云龙, 项韬, 朱猛进, 余梅, 李新云, 刘小磊, 赵书红. 全基因组选择模型研究进展及展望[J]. 畜牧兽医学报, 2019, 50(2): 233-242.
[14]	王延晖,朱波，李俊雅. 基于显性模型的基因组选择中贝叶斯方法研究[J]. 畜牧兽医学报, 2017, 48(1): 60-67.
[15]	王晨，秦珂，薛明，莫德林，陈瑶生，刘小红. 全基因组选择在猪育种中的应用[J]. 畜牧兽医学报, 2016, 47(1): 1-9.