DOI QR코드

DOI QR Code

Application of deep learning with bivariate models for genomic prediction of sow lifetime productivity-related traits

  • Joon-Ki Hong (Swine Division, National Institute of Animal Science, Rural Development Administration) ;
  • Yong-Min Kim (Swine Division, National Institute of Animal Science, Rural Development Administration) ;
  • Eun-Seok Cho (Swine Division, National Institute of Animal Science, Rural Development Administration) ;
  • Jae-Bong Lee (Korea Zoonosis Research Institute, Jeonbuk National University) ;
  • Young-Sin Kim (Swine Division, National Institute of Animal Science, Rural Development Administration) ;
  • Hee-Bok Park (Department of Animal Resources Science, Kongju National University)
  • 투고 : 2023.07.14
  • 심사 : 2023.11.03
  • 발행 : 2024.04.01

초록

Objective: Pig breeders cannot obtain phenotypic information at the time of selection for sow lifetime productivity (SLP). They would benefit from obtaining genetic information of candidate sows. Genomic data interpreted using deep learning (DL) techniques could contribute to the genetic improvement of SLP to maximize farm profitability because DL models capture nonlinear genetic effects such as dominance and epistasis more efficiently than conventional genomic prediction methods based on linear models. This study aimed to investigate the usefulness of DL for the genomic prediction of two SLP-related traits; lifetime number of litters (LNL) and lifetime pig production (LPP). Methods: Two bivariate DL models, convolutional neural network (CNN) and local convolutional neural network (LCNN), were compared with conventional bivariate linear models (i.e., genomic best linear unbiased prediction, Bayesian ridge regression, Bayes A, and Bayes B). Phenotype and pedigree data were collected from 40,011 sows that had husbandry records. Among these, 3,652 pigs were genotyped using the PorcineSNP60K BeadChip. Results: The best predictive correlation for LNL was obtained with CNN (0.28), followed by LCNN (0.26) and conventional linear models (approximately 0.21). For LPP, the best predictive correlation was also obtained with CNN (0.29), followed by LCNN (0.27) and conventional linear models (approximately 0.25). A similar trend was observed with the mean squared error of prediction for the SLP traits. Conclusion: This study provides an example of a CNN that can outperform against the linear model-based genomic prediction approaches when the nonlinear interaction components are important because LNL and LPP exhibited strong epistatic interaction components. Additionally, our results suggest that applying bivariate DL models could also contribute to the prediction accuracy by utilizing the genetic correlation between LNL and LPP.

키워드

과제정보

The authors sincerely thank Miguel Perez-Enciso, at Universitat Autonoma de Barcelona (UAB), for valuable comments on the manuscript.

참고문헌

  1. Sasaki Y, Koketsu Y. Sows having high lifetime efficiency and high longevity associated with herd productivity in commercial herds. Livest Sci 2008;118:140-6. https://doi.org/10.1016/j.livsci.2007.12.029
  2. Onteru SK, Fan B, Nikkila MT, Garrick DJ, Stalder KJ, Rothschild MF. Whole-genome association analyses for lifetime reproductive traits in the pig. J Anim Sci 2011;89:988-95. https://doi.org/10.2527/jas.2010-3236
  3. Iida R, Pineiro C, Koketsu Y. High lifetime and reproductive performance of sows on southern European Union commercial farms can be predicted by high numbers of pigs born alive in parity one. J Anim Sci 2015;93:2501-8. https://doi.org/10.2527/jas.2014-8781
  4. Sevon-Aimonen M, Uimari P. Heritability of sow longevity and lifetime prolificacy in Finnish Yorkshire and Landrace pigs. Agric Food Sci 2013;22:325-30. https://doi.org/10.23986/afsci.7991
  5. Serenius T, Stalder KJ. Length of productive life of crossbred sows is affected by farm management, leg conformation, sow's own prolificacy, sow's origin parity and genetics. Animal 2007;1:745-50. https://doi.org/10.1017/S175173110769185X
  6. Noppibool U, Koonawootrittriron S, Elzo MA, Suwanasopee T. Factors affecting length of productive life and lifetime production traits in a commercial swine herd in Northern Thailand. Agric Nat Resour 2016;50:71-4. https://doi.org/10.1016/j.anres.2015.07.001
  7. Paixao G, Martins A, Esteves A, Payan-Carreira R, Carolino N. Genetic parameters for reproductive, longevity and lifetime production traits in Bisaro pigs. Livest Sci 2019;225:129-34. https://doi.org/10.1016/j.livsci.2019.05.010
  8. Meuwissen T, Hayes B, Goddard M. Accelerating improvement of livestock with genomic selection. Annu Rev Anim Biosci 2013;1:221-37. https://doi.org/10.1146/annurev-animal-031412-103705
  9. de Los Campos G, Hickey JM, Pong-Wong R, Daetwyler HD, Calus MPL. Whole-genome regression and prediction methods applied to plant and animal breeding. Genetics 2013;193:327-45. https://doi.org/10.1534/genetics.112.143313
  10. Gianola D, Rosa GJM. One hundred years of statistical developments in animal breeding. Annu Rev Anim Biosci 2015;3:19-56. https://doi.org/10.1146/annurev-animal-022114-110733
  11. Calus MPL, Veerkamp RF. Accuracy of multi-trait genomic selection using different methods. Genet Sel Evol 2011;43:26. https://doi.org/10.1186/1297-9686-43-26
  12. Cheng H, Kizilkaya K, Zeng J, Garrick D, Fernando R. Genomic prediction from multiple-trait Bayesian regression methods using mixture priors. Genetics 2018;209:89-103. https://doi.org/10.1534/genetics.118.300650
  13. Mehrban H, Naserkheil M, Lee D, Ibanez-Escriche N. Multitrait single-step GBLUP improves accuracy of genomic prediction for carcass traits using yearling weight and ultrasound traits in Hanwoo. Front Genet 2021;12:692356. https://doi.org/10.3389/fgene.2021.692356
  14. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015;521:436-44. https://doi.org/10.1038/nature14539
  15. Eraslan G, Avsec Z, Gagneur J, Theis FJ. Deep learning: new computational modelling techniques for genomics. Nat Rev Genet 2019;20:389-403. https://doi.org/10.1038/s41576-019-0122-6
  16. Kasani PH, Oh SM, Choi YH, et al. A computer vision-based approach for behavior recognition of gestating sows fed different fiber levels during high ambient temperature. J Anim Sci Technol 2021;63:367-79. https://doi.org/10.5187/jast.2021.e35
  17. Kim M, Choi Y, Lee J, Sa S, Cho H. A deep learning-based approach for feeding behavior recognition of weanling pigs. J Anim Sci Technol 2021;63:1453-63. https://doi.org/10.5187/jast.2021.e127
  18. Montesinos-Lopez OA, Montesinos-Lopez A, Perez-Rodriguez P, et al. A review of deep learning applications for genomic selection. BMC Genomics 2021;22:19. https://doi.org/10.1186/s12864-020-07319-x
  19. Perez-Enciso M, Zingaretti LM. A guide on deep learning for complex trait genomic prediction. Genes 2019;10:553. https://doi.org/10.3390/genes10070553
  20. Bellot P, de los Campos G, Perez-Enciso M. Can deep learning improve genomic prediction of complex human traits? Genetics 2018;210:809-19. https://doi.org/10.1534/genetics.118.301298
  21. Waldmann P. Approximate Bayesian neural networks in genomic prediction. Genet Sel Evol 2018;50:70. https://doi.org/10.1186/s12711-018-0439-1
  22. Abdollahi-Arpanahi R, Gianola D, Penagaricano F. Deep learning versus parametric and ensemble methods for genomic prediction of complex phenotypes. Genet Sel Evol 2020;52:12. https://doi.org/10.1186/s12711-020-00531-z
  23. Montesinos-Lopez OA, Montesinos-Lopez A, Crossa J, Gianola D, Hernandez-Suarez CM, Martin-Vallejo J. Multi-trait, multi-environment deep learning modeling for genomic-enabled prediction of plant traits. G3-Genes Genomes Genet 2018;8:3829-40. https://doi.org/10.1534/g3.118.200728
  24. Montesinos-Lopez OA, Martin-Vallejo J, Crossa J, et al. New deep learning genomic-based prediction model for multiple traits with binary, ordinal, and continuous phenotypes. G3-Genes Genomes Genet 2019;9:1545-56. https://doi.org/10.1534/g3.119.300585
  25. Sandhu K, Patil SS, Pumphrey M, Carter A. Multitrait machineand deep-learning models for genomic selection using spectral information in a wheat breeding program. Plant Genome 2021;14:e20119. https://doi.org/https://doi.org/10.1002/tpg2.20119
  26. Perez-Rodriguez P, Gianola D, Gonzalez-Camacho JM, Crossa J, Manes Y, Dreisigacker S. Comparison between linear and non-parametric regression models for genomeenabled prediction in wheat. G3-Genes Genomes Genet 2012;2:1595-605. https://doi.org/10.1534/g3.112.003665
  27. Gonzalez-Camacho JM, Crossa J, Perez-Rodriguez P, Ornella L, Gianola D. Genome-enabled prediction using probabilistic neural network classifiers. BMC Genomics 2016;17:208. https://doi.org/10.1186/s12864-016-2553-1
  28. Ma W, Qiu Z, Song J, et al. A deep convolutional neural network approach for predicting phenotypes from genotypes. Planta 2018;248:1307-18. https://doi.org/10.1007/s00425-018-2976-9
  29. Montesinos-Lopez OA, Montesinos-Lopez JC, Singh P, et al. A multivariate poisson deep learning model for genomic prediction of count data. G3-Genes Genomes Genet 2020;10:4177-90. https://doi.org/10.1534/g3.120.401631
  30. Zingaretti LM, Gezan SA, Ferrao LFV, et al. Exploring deep learning for complex trait genomic prediction in polyploid outcrossing species. Front Plant Sci 2020;11:25. https://doi.org/10.3389/fpls.2020.00025
  31. Waldmann P, Pfeiffer C, Meszaros G. Sparse convolutional neural networks for genome-wide prediction. Front Genet 2020;11:25. https://doi.org/10.3389/fgene.2020.00025
  32. Alzubaidi L, Zhang J, Humaidi AJ, et al. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. J Big Data 2021;8:53. https://doi.org/10.1186/s40537-021-00444-8
  33. Pook T, Freudenthal J, Korte A, Simianer H. Using local convolutional neural networks for genomic prediction. Front Genet 2020;11:561497. https://doi.org/10.3389/fgene.2020.561497
  34. Choy YH, Mahboob A, Cho CI, et al. Genetic parameters of pre-adjusted body weight growth and ultrasound measures of body tissue development in three seedstock pig breed populations in Korea. Asian-Australas J Anim Sci 2015;28:1696-702. https://doi.org/10.5713/ajas.14.0971
  35. Alam M, Chang HK, Lee SS, Choi TJ. Genetic analysis of major production and reproduction traits of Korean Duroc, Landrace and Yorkshire pigs. Animals 2021;11:1321. https://doi.org/10.3390/ani11051321
  36. Gozalo-Marcilla M, Buntjer J, Johnsson M, et al. Genetic architecture and major genes for backfat thickness in pig lines of diverse genetic backgrounds. Genet Sel Evol 2021;53:76. https://doi.org/10.1186/s12711-021-00671-w
  37. Sargolzaei M, Chesnais JP, Schenkel FS. A new approach for efficient genotype imputation using information from relatives. BMC Genomics 2014;15:478. https://doi.org/10.1186/1471-2164-15-478
  38. VanRaden PM. Efficient methods to compute genomic predictions. J Dairy Sci 2008;91:4414-23. https://doi.org/10.3168/jds.2007-0980
  39. Su G, Christensen OF, Ostersen T, Henryon M, Lund MS. Estimating additive and non-additive genetic variances and predicting genetic merits using genome-wide dense single nucleotide polymorphism markers. PLoS One 2012;7:e45293. https://doi.org/10.1371/journal.pone.0045293
  40. Vitezica ZG, Legarra A, Toro MA, Varona L. Orthogonal estimates of variances for additive, dominance, and epistatic effects in populations. Genetics 2017;206:1297-307. https://doi.org/10.1534/genetics.116.199406
  41. Perez P, de los Campos G. Genome-wide regression and prediction with the BGLR statistical package. Genetics 2014; 198:483-95. https://doi.org/10.1534/genetics.114.164442
  42. Montesinos-Lopez OA, Montesinos-Lopez A, Crossa J, et al. A singular value decomposition Bayesian multiple-trait and multiple-environment genomic model. Heredity 2019;122: 381-401. https://doi.org/10.1038/s41437-018-0109-7
  43. Gianola D. Priors in whole-genome regression: the bayesian alphabet returns. Genetics 2013;194:573-96. https://doi.org/10.1534/genetics.113.151753
  44. Chollet F, Allaire J, Falbel D, Tang Y, Van Der Bijl W, Studer M. R interface to keras. Keras Team c2017 [cited 2020, Apr 21]. Available from: https://github.com/rstudio/keras
  45. Kingma DP, Ba J. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 2014. https://doi.org/10.48550/arXiv.1412.6980
  46. Yan Y. rBayesianOptimization: Bayesian optimization of hyperparameters. R package version; 2016. Available from: https://github.com/yanyachen/rBayesianOptimization
  47. Vitezica ZG, Reverter A, Herring W, Legarra A. Dominance and epistatic genetic variances for litter size in pigs using genomic models. Genet Sel Evol 2018;50:71. https://doi.org/10.1186/s12711-018-0437-3
  48. Jia Y, Jannink J. Multiple-trait genomic selection methods increase genetic value prediction accuracy. Genetics 2012;192:1513-22. https://doi.org/10.1534/genetics.112.144246
  49. Lancaster J, Lorenz R, Leech R, Cole JH. Bayesian optimization for neuroimaging pre-processing in brain age classification and prediction. Front Aging Neurosci 2018;10;28. https://doi.org/10.3389/fnagi.2018.00028