{"title":"Comparison of predictive ability of single-trait and multitrait genomic selection models for body growth traits in Maiwa yaks","authors":"","doi":"10.1016/j.animal.2024.101350","DOIUrl":null,"url":null,"abstract":"<div><div>Yaks are grazed extensively on the Qinghai–Tibet Plateau, which has a long history of semi-domestication. The predicted weight of yaks over consecutive years helps make strategic decisions when selecting yak calves for breeding. To achieve more accurate predictions of genomic estimated breeding values, we used a dataset comprising the genotype and weight records of 396 Maiwa yaks collected from 2015 to 2020. We compared the predictive accuracy of the genome best linear unbiased prediction (<strong>GBLUP</strong>) model with that of six other models. Based on the GBLUP model, we applied two prediction strategies. In the first strategy, the year was treated as a fixed effect in the GBLUP model, and the kinship from all individuals and the markers were treated as random effects. In the second strategy, all individuals were divided into six age groups, with GBLUP performed for each group, and the phenotypes of the closest age groups were treated as fixed effects. Although the GBLUP model provided better prediction accuracy than other single-trait models, most of the predictive capacity was derived from the best linear unbiased estimation. Additionally, incorporating the phenotype of the closest age group as a factor in multitrait prediction enhanced the accuracy of the model. Our findings provide a robust and credible strategy for predicting continuous economic traits in the presence of strong correlations.</div></div>","PeriodicalId":50789,"journal":{"name":"Animal","volume":null,"pages":null},"PeriodicalIF":4.0000,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Animal","FirstCategoryId":"97","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1751731124002878","RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"AGRICULTURE, DAIRY & ANIMAL SCIENCE","Score":null,"Total":0}
引用次数: 0
Abstract
Yaks are grazed extensively on the Qinghai–Tibet Plateau, which has a long history of semi-domestication. The predicted weight of yaks over consecutive years helps make strategic decisions when selecting yak calves for breeding. To achieve more accurate predictions of genomic estimated breeding values, we used a dataset comprising the genotype and weight records of 396 Maiwa yaks collected from 2015 to 2020. We compared the predictive accuracy of the genome best linear unbiased prediction (GBLUP) model with that of six other models. Based on the GBLUP model, we applied two prediction strategies. In the first strategy, the year was treated as a fixed effect in the GBLUP model, and the kinship from all individuals and the markers were treated as random effects. In the second strategy, all individuals were divided into six age groups, with GBLUP performed for each group, and the phenotypes of the closest age groups were treated as fixed effects. Although the GBLUP model provided better prediction accuracy than other single-trait models, most of the predictive capacity was derived from the best linear unbiased estimation. Additionally, incorporating the phenotype of the closest age group as a factor in multitrait prediction enhanced the accuracy of the model. Our findings provide a robust and credible strategy for predicting continuous economic traits in the presence of strong correlations.
期刊介绍:
Editorial board
animal attracts the best research in animal biology and animal systems from across the spectrum of the agricultural, biomedical, and environmental sciences. It is the central element in an exciting collaboration between the British Society of Animal Science (BSAS), Institut National de la Recherche Agronomique (INRA) and the European Federation of Animal Science (EAAP) and represents a merging of three scientific journals: Animal Science; Animal Research; Reproduction, Nutrition, Development. animal publishes original cutting-edge research, ''hot'' topics and horizon-scanning reviews on animal-related aspects of the life sciences at the molecular, cellular, organ, whole animal and production system levels. The main subject areas include: breeding and genetics; nutrition; physiology and functional biology of systems; behaviour, health and welfare; farming systems, environmental impact and climate change; product quality, human health and well-being. Animal models and papers dealing with the integration of research between these topics and their impact on the environment and people are particularly welcome.