1
|
Pedrosa VB, Chen SY, Gloria LS, Doucette JS, Boerman JP, Rosa GJM, Brito LF. Machine learning methods for genomic prediction of cow behavioral traits measured by automatic milking systems in North American Holstein cattle. J Dairy Sci 2024:S0022-0302(24)00497-1. [PMID: 38395400 DOI: 10.3168/jds.2023-24082] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/12/2023] [Accepted: 01/18/2024] [Indexed: 02/25/2024]
Abstract
Identifying genome-enabled methods that provide more accurate genomic prediction is crucial when evaluating complex traits such as dairy cow behavior. In this study, we aimed to compare the predictive performance of traditional genomic prediction methods and deep learning algorithms for genomic prediction of milking refusals (MREF) and milking failures (MFAIL) in North American Holstein cows measured by automatic milking systems (milking robots). A total of 1,993,509 daily records from 4,511 genotyped Holstein cows were collected by 36 milking robot stations. After quality control, 57,600 single nucleotide polymorphisms (SNP) were available for the analyses. Four genomic prediction methods were considered: Bayesian Lasso (LASSO), Multiple Layer Perceptron (MLP), Convolutional Neural Network (CNN), and Genomic Best Linear Unbiased Prediction (GBLUP). We implemented the first 3 methods using the Keras and TensorFlow libraries in Python (v.3.9) while the GBLUP method was implemented using the BLUPF90+ family programs. The accuracy of genomic prediction (Mean Square Error) for MREF and MFAIL was 0.34 (0.08) and 0.27 (0.08) based on LASSO, 0.36 (0.09) and 0.32 (0.09) for MLP, 0.37 (0.08) and 0.30 (0.09) for CNN, and 0.35 (0.09) and 0.31(0.09) based on GBLUP, respectively. Additionally, we observed a lower re-ranking of top selected individuals based on the MLP versus CNN methods compared with the other approaches for both MREF and MFAIL. Although the deep learning methods showed slightly higher accuracies than GBLUP, the results may not be sufficient to justify their use over traditional methods due to their higher computational demand and the difficulty of performing genomic prediction for non-genotyped individuals using deep learning procedures. Overall, this study provides insights into the potential feasibility of using deep learning methods to enhance genomic prediction accuracy for behavioral traits in livestock. Further research is needed to determine their practical applicability to large dairy cattle breeding programs.
Collapse
Affiliation(s)
- Victor B Pedrosa
- Department of Animal Sciences, Purdue University, West Lafayette, IN, 47907, USA
| | - Shi-Yi Chen
- Department of Animal Sciences, Purdue University, West Lafayette, IN, 47907, USA; Farm Animal Genetic Resources Exploration and Innovation Key Laboratory of Sichuan Province, Sichuan Agricultural University, Chengdu, Sichuan, 611130, China
| | - Leonardo S Gloria
- Department of Animal Sciences, Purdue University, West Lafayette, IN, 47907, USA
| | - Jarrod S Doucette
- Agriculture Information Technology (AgIT), Purdue University, West Lafayette, IN, 47907, USA
| | - Jacquelyn P Boerman
- Department of Animal Sciences, Purdue University, West Lafayette, IN, 47907, USA
| | - Guilherme J M Rosa
- Department of Animal and Dairy Sciences, University of Wisconsin-Madison, Madison, WI, 53706, USA
| | - Luiz F Brito
- Department of Animal Sciences, Purdue University, West Lafayette, IN, 47907, USA.
| |
Collapse
|
2
|
Zhuo Y, Du H, Diao C, Li W, Zhou L, Jiang L, Jiang J, Liu J. MAGE: metafounders-assisted genomic estimation of breeding value, a novel additive-dominance single-step model in crossbreeding systems. Bioinformatics 2024; 40:btae044. [PMID: 38268487 DOI: 10.1093/bioinformatics/btae044] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/23/2023] [Revised: 01/07/2024] [Accepted: 01/22/2024] [Indexed: 01/26/2024] Open
Abstract
MOTIVATION Utilizing both purebred and crossbred data in animal genetics is widely recognized as an optimal strategy for enhancing the predictive accuracy of breeding values. Practically, the different genetic background among several purebred populations and their crossbred offspring populations limits the application of traditional prediction methods. Several studies endeavor to predict the crossbred performance via the partial relationship, which divides the data into distinct sub-populations based on the common genetic background, such as one single purebred population and its corresponding crossbred descendant. However, this strategy makes prediction inaccurate due to ignoring half of the parental information of crossbreed animals. Furthermore, dominance effects, although playing a significant role in crossbreeding systems, cannot be modeled under such a prediction model. RESULTS To overcome this weakness, we developed a novel multi-breed single-step model using metafounders to assess ancestral relationships across diverse breeds under a unified framework. We proposed to use multi-breed dominance combined relationship matrices to model additive and dominance effects simultaneously. Our method provides a straightforward way to evaluate the heterosis of crossbreeds and the breeding values of purebred parents efficiently and accurately. We performed simulation and real data analyses to verify the potential of our proposed method. Our proposed model improved prediction accuracy under all scenarios considered compared to commonly used methods. AVAILABILITY AND IMPLEMENTATION The software for implementing our method is available at https://github.com/CAU-TeamLiuJF/MAGE.
Collapse
Affiliation(s)
- Yue Zhuo
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - Heng Du
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - ChenGuang Diao
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - WeiNing Li
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - Lei Zhou
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - Li Jiang
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| | - JiCai Jiang
- Department of Animal Science, North Carolina State University, Raleigh, NC 27695, United States
| | - JianFeng Liu
- State Key Laboratory of Animal Biotech Breeding, College of Animal Science and Technology, China Agricultural University, Beijing 100193, China
| |
Collapse
|
3
|
Zhao W, Qadri QR, Zhang Z, Wang Z, Pan Y, Wang Q, Zhang Z. PyAGH: a python package to fast construct kinship matrices based on different levels of omic data. BMC Bioinformatics 2023; 24:153. [PMID: 37072709 PMCID: PMC10111838 DOI: 10.1186/s12859-023-05280-6] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/03/2022] [Accepted: 04/10/2023] [Indexed: 04/20/2023] Open
Abstract
BACKGROUND Construction of kinship matrices among individuals is an important step for both association studies and prediction studies based on different levels of omic data. Methods for constructing kinship matrices are becoming diverse and different methods have their specific appropriate scenes. However, software that can comprehensively calculate kinship matrices for a variety of scenarios is still in an urgent demand. RESULTS In this study, we developed an efficient and user-friendly python module, PyAGH, that can accomplish (1) conventional additive kinship matrces construction based on pedigree, genotypes, abundance data from transcriptome or microbiome; (2) genomic kinship matrices construction in combined population; (3) dominant and epistatic effects kinship matrices construction; (4) pedigree selection, tracing, detection and visualization; (5) visualization of cluster, heatmap and PCA analysis based on kinship matrices. The output from PyAGH can be easily integrated in other mainstream software based on users' purposes. Compared with other softwares, PyAGH integrates multiple methods for calculating the kinship matrix and has advantages in terms of speed and data size compared to other software. PyAGH is developed in python and C + + and can be easily installed by pip tool. Installation instructions and a manual document can be freely available from https://github.com/zhaow-01/PyAGH . CONCLUSION PyAGH is a fast and user-friendly Python package for calculating kinship matrices using pedigree, genotype, microbiome and transcriptome data as well as processing, analyzing and visualizing data and results. This package makes it easier to perform predictions and association studies processes based on different levels of omic data.
Collapse
Affiliation(s)
- Wei Zhao
- Department of Animal Science, School of Agriculture and Biology, Shanghai Jiao Tong University, 800# Dongchuan Road, Shanghai, China
| | - Qamar Raza Qadri
- Department of Animal Science, School of Agriculture and Biology, Shanghai Jiao Tong University, 800# Dongchuan Road, Shanghai, China
| | - Zhenyang Zhang
- Department of Animal Science, College of Animal Sciences, Zhejiang University, 866# Yuhangtang Road, Hangzhou, 310058, China
| | - Zhen Wang
- Department of Animal Science, College of Animal Sciences, Zhejiang University, 866# Yuhangtang Road, Hangzhou, 310058, China
| | - Yuchun Pan
- Department of Animal Science, College of Animal Sciences, Zhejiang University, 866# Yuhangtang Road, Hangzhou, 310058, China
- Hainan Research Institute, Zhejiang University, 11# Yonyou Industrial Park, Yazhou Bay Science and Technology City, Sanya, 572025, China
| | - Qishan Wang
- Department of Animal Science, College of Animal Sciences, Zhejiang University, 866# Yuhangtang Road, Hangzhou, 310058, China.
| | - Zhe Zhang
- Department of Animal Science, College of Animal Sciences, Zhejiang University, 866# Yuhangtang Road, Hangzhou, 310058, China.
| |
Collapse
|