Logo do repositório
 

Estratégias para predição genômica em populações multirraciais de zebuínos aplicando o método do passo único genômico BLUP

dc.contributor.advisorBaldi, Fernando Sebastian [UNESP]
dc.contributor.authorLondoño-Gil, Marisol [UNESP]
dc.contributor.coadvisorLourenço, Daniela Andressa Lino
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)pt
dc.date.accessioned2025-05-21T16:19:53Z
dc.date.issued2025-05-09
dc.description.abstractA seleção genômica transformou a seleção de animais ao permitir a identificação precoce de indivíduos superiores com base em seu mérito genético. Contudo, essa abordagem enfrenta desafios consideráveis quando aplicada ao gado zebuíno, especialmente em populações como Nelore, Brahman, Guzerá e Tabapuã. Brahman, Guzerá e Tabapuã apresentam populações de referência pequenas, o que torna a sua seleção genômica difícil e imprecisa. Uma avaliação multirracial poderia resolver esse problema, mas deve considerar a heterogeneidade nas frequências alélicas que podem existir entre as raças e a incompatibilidade entre as matrizes de relacionamento genômico (G) e de pedigree (A). Este trabalho pesquisou estratégias para as predições genômicas de gado zebuíno, em uma população multirracial que incluiu o Nelore, Brahman, Guzerá e Tabapuã, utilizando o preditor linear não viesado genômico de uma única etapa (ssGBLUP). Foram utilizados dados de aproximadamente 4,2 milhões de registros de pedigree, 329 mil observações fenotípicas e 63 mil animais genotipados. As características analisadas foram o peso ajustado aos 210 dias (W210), o peso ajustado aos 450 dias (W450) e a circunferência escrotal aos 365 dias (SC365). Diferentes cenários foram avaliados, comparando abordagens de raça única e multirracial. As avaliações multirraciais foram conduzidas tanto com a inclusão de metafundadores (MF), para levar em conta as diferenças e similitudes genéticas entre as raças, quanto sem eles. Sem MF, foi analisada o ajuste ou não de frequências alélicas específicas de cada raça. Análises comparativas dos diferentes cenários foram realizadas utilizando o método de regressão linear (LR) para avaliar viés, dispersão e acurácia. Além das avaliações genômicas diretas, este trabalho abordou o desafio de obter predições genômicas para animais jovens genotipados que carecem fenótipos ou pedigree, calculando predições indiretas (IP). Essas IP foram derivadas como a soma dos efeitos dos SNPs ponderados pelo conteúdo gênico. A avaliação dos diferentes cenários foi realizada utilizando o método LR. Ademais, a acurácia das IP (ACCIP) foi estimada a partir da covariância do erro de predição (PEC) dos efeitos dos SNPs e validada através de correlações com a acurácia dos valores genômicos estimados (ACCGEBV). Os resultados deste trabalho demonstraram que as avaliações multirraciais melhoraram significativamente a precisão das predições em comparação com as abordagens de raça única, beneficiando particularmente as raças com conjuntos de dados genômicos menores, como Guzerá e Tabapuã. A inclusão de MF reduziu o viés e mitigou a sub ou superdispersão nas predições, com um modesto aumento na acurácia das IP. As avaliações de raça única frequentemente apresentaram menor precisão, especialmente para Guzerá e Tabapuã, que possuem dados genômicos limitados. O estudo demonstrou que as avaliações genômicas multirraciais estruturadas são uma estratégia robusta e eficaz para melhorar a seleção genômica no gado zebuíno brasileiro. Ao utilizar informações genômicas de múltiplas raças, essa abordagem melhorou a precisão das predições para características de importância econômica e permitiu a rápida obtenção de IP para animais jovens. Consequentemente, as avaliações multirraciais de gado zebuíno brasileiro podem auxiliar nas decisões de seleção e potencialmente acelerar o progresso genético, especialmente em raças com populações de referência pequenas.pt
dc.description.abstractGenomic selection has changed animal breeding by enabling the early identification of superior individuals based on their genetic merit. However, this approach faces considerable challenges when applied to indicine cattle, particularly populations such as Nellore, Brahman, Guzerat, and Tabapua. Brahman, Guzerat, and Tabapua present small reference populations, making genomic selection difficult and inaccurate. A multi-breed evaluation could solve this problem, but it must consider the heterogeneity in allele frequencies among breeds and the incompatibility between genomic (G) and pedigree (A) relationship matrices. This work investigated strategies, feasibility, and effectiveness for genomic prediction in indicine cattle by focusing on a multi‐breed population that includes Nellore, Brahman, Guzerat, and Tabapua using the single‐step genomic best linear unbiased predictor (ssGBLUP). Data for approximately 4.2 M pedigree records, 329 K phenotypic observations, and 63 K genotyped animals were used. The traits analyzed were adjusted weight at 210 days (W210), adjusted weight at 450 days (W450), and scrotal circumference at 365 days (SC365). Different scenarios were evaluated, comparing single-breed and multi-breed approaches. Multi-breed evaluations were conducted both with and without the inclusion of metafounders (MF) to account for genetic differences across breeds. At the same time, without MF, the inclusion of breed-specific allele frequencies was analyzed or not adjusted at all. Comparative analyses of the different scenarios were performed using the linear regression (LR) method to assess bias, dispersion, and accuracy. In addition to improving direct genomic evaluations, this work addressed the challenge of obtaining genomic predictions for young, genotyped animals lacking phenotypic data or pedigree information by computing indirect predictions (IP). These IP were derived as the sum of SNP effects weighted by gene content. The evaluation of the different scenarios was made using the LR method. Moreover, the accuracy of IP (ACCIP) was estimated from the prediction error covariance (PEC) of SNP effects and validated through correlations with the accuracy of genomic estimated breeding values (ACCGEBV). The results of this work showed that multi-breed evaluations significantly improved prediction accuracy compared to single-breed approaches, particularly helping breeds with smaller genomic datasets such as Guzerat, and Tabapua. Incorporating MF reduced bias and mitigated under- or over-dispersion in the predictions, with a modest increase in IP accuracy. The single-breed evaluations frequently showed lower accuracy, especially for Guzerat and Tabapua which have limited genomic data. The study demonstrated that structured multi-breed genomic evaluations are a robust and effective strategy for improving genomic selection in Brazilian indicine cattle. While using genomic information across multiple breeds, this approach improved the accuracy of predictions for traits of economic importance and allowed rapid IP for young animals. Consequently, Brazilian indicine multi-breed evaluations can help selection decisions and potentially accelerate genetic progress, especially in breeds with constrained reference populations.en
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.description.sponsorshipIdFAPESP: 2020/14275-5
dc.description.sponsorshipIdFAPESP: 2022/15168-3
dc.description.sponsorshipIdFAPESP: 2021/04807-2
dc.description.sponsorshipIdCNPq: 402313/2021-6
dc.identifier.capes33004102002P0
dc.identifier.citationLONDOÑO-GIL, M. - Estratégias para predição genômica em populações multirraciais de zebuínos aplicando o método do passo único genômico BLUP - 2025, 136f - Tese (Doutorado em Ciência Animal) - Universidade Estadual Paulista "Júlio de Mesquita Filho", Jaboticabal, 2025.pt
dc.identifier.lattes7973071842374266
dc.identifier.orcidhttps://orcid.org/0000-0001-6522-5567
dc.identifier.urihttps://hdl.handle.net/11449/310567
dc.language.isoeng
dc.publisherUniversidade Estadual Paulista (Unesp)
dc.rights.accessRightsAcesso restritopt
dc.subjectGuzeráen
dc.subjectAnimais melhoramento genéticoen
dc.subjectGenéticapt
dc.subjectTabapuapt
dc.titleEstratégias para predição genômica em populações multirraciais de zebuínos aplicando o método do passo único genômico BLUPpt
dc.title.alternativeStrategies for genomic predictions of an Indicine multibreed population using singlestep gblupen
dc.typeTese de doutoradopt
dcterms.impactBrazil is home to one of the largest cattle populations in the world, with more than 238 million animals spread across the country. Among the most important zebu breeds are Nellore, Guzerat, Tabapua, and Brahman. These breeds are well adapted to tropical climates and perform well even under challenging pasture conditions. However, when it comes to applying genomic tools, that is, predicting an animal’s potential based on its DNA, there is a significant imbalance. Nellore has a big amount of genomic and performance data available, while the other breeds lag far behind, limiting the accuracy and reliability of genetic evaluations for young animals that could become the next generation of sires. This research addresses that gap by proposing a multi-breed genomic evaluation strategy. Instead of evaluating each breed separately, the study integrates data from all four breeds into a single, unified evaluation model. Using a method known as metafounders, the genetic differences between breeds are respected and accounted for, while still allowing the information to be shared across populations. This makes genetic predictions more accurate and stable, especially for underrepresented breeds like Guzerat and Tabapua. Another major contribution of this work is the use of indirect predictions (IP), a method that estimates the genetic merit of young animals without performance or pedigree records. These predictions are calculated using information from older, wellevaluated animals and allow for early selection decisions. This is particularly valuable in commercial herds, where rapid and cost-effective decision-making is critical. IP enables producers to identify promising animals at an earlier age, reducing feeding and management costs, and accelerating the pace of genetic improvement. Moreover, the method simplifies the use of genomic selection in routine evaluations. Since IP relies on previously estimated SNP effects, there is no need to recalculate entire genomic evaluations each time a new animal is genotyped. This greatly reduces computational costs and makes genomic tools more accessible to a wider range of breeders. The real-world impact of this research is already visible. The 2024 Sire Report published by ANCP (National Association of Breeders and Researchers) officially adopted the multi-breed evaluation strategy developed in this study. This has given breeders access to more reliable and science-based tools for selection, even for breeds that previously lacked robust genomic information. Furthermore, countries that collaborate with ANCP and work with Brahman cattle, for example, can now use indirect predictions for selection, even if they don’t yet have performance data, as long as genomic data is available. Looking ahead, this research has the potential to transform how genetic improvement is approached in tropical cattle systems worldwide. As more countries and organizations adopt multi-breed evaluations and indirect prediction strategies, we can expect a wider and fairer distribution of the benefits of genomics, particularly in low- and middle-income countries where data collection remains a challenge. The methodology developed here also opens the door to international genetic evaluations between indicine breeds, promoting genetic exchange and collaboration across countries. This could increase the genetic diversity and resilience of cattle populations, supporting long-term sustainability in the face of climate change and evolving market demands. Additionally, by enabling faster and earlier selection, this approach can significantly reduce the generation interval, making it possible to achieve genetic gains more quickly and efficiently. This is a game changer not just for elite breeding programs, but also for small and medium producers who can now access high-impact genetic tools without needing complex infrastructure. Ultimately, this research provides a scalable and inclusive framework for implementing genomic selection in real-world settings. As technology becomes more accessible and costs continue to drop, the strategies proposed here can serve as a model for other livestock species and agricultural sectors aiming to increase productivity, sustainability, and food security through science-based innovation.en
dspace.entity.typePublication
unesp.campusUniversidade Estadual Paulista (UNESP), Faculdade de Ciências Agrárias e Veterinárias, Jaboticabalpt
unesp.embargo12 meses após a data da defesapt
unesp.examinationboard.typeBanca públicapt
unesp.graduateProgramCiência Animal - FCAVpt
unesp.knowledgeAreaAgriculturapt
unesp.researchAreaGenética e Melhoramento Animalpt

Arquivos

Pacote original

Agora exibindo 1 - 2 de 2
Carregando...
Imagem de Miniatura
Nome:
londoño-gil_m_dr_jabo_int_sub.pdf
Tamanho:
4.35 MB
Formato:
Adobe Portable Document Format
Carregando...
Imagem de Miniatura
Nome:
londoño-gil_m_dr_jabo_par_sub.pdf
Tamanho:
1.95 MB
Formato:
Adobe Portable Document Format

Licença do pacote

Agora exibindo 1 - 1 de 1
Carregando...
Imagem de Miniatura
Nome:
license.txt
Tamanho:
2.14 KB
Formato:
Item-specific license agreed upon to submission
Descrição: