Causal networks, genomic prediction and candidate genes for boar taint compounds

• Ano de defesa: 2020
• Autor(a) principal: Botelho, Margareth Evangelista
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Viçosa; Zootecnia
• Palavras-chave em Português: Suínos - Genética; Androstenona; Escatol; Indol; Genômica; Genes; Genética e Melhoramento dos Animais Domésticos
• Link de acesso: https://locus.ufv.br//handle/123456789/29403

The piglet non-castration may result in the boar taint appearance, which is an unpleasant taste and smell in pig meat. Boar taint is caused by the increasing of boar taint compounds (androstenone, skatole and indole) levels in adipose tissue. However, the genetic architecture and the causal relationship between levels of boar taint compounds in adipose tissue still require elucidation. In this sense, firstly, we studied the causal relationship between androstenone, skatole and indole levels in carcass adipose tissue samples and animal biopsies using structural equations models (SEM). In summary, we verified that using priori information to define the causal structure increased the model goodness-of-fit, however, the credibility intervals were also increased resulting in several unexpected null genetic correlation. We identified direct and indirect effects between boar taint compounds, mainly androstenone in biopsies affect skatole in carcass and skatole in carcass affect androstenone in carcass. Posteriorly, we evaluated the effects of SNPs weighting strategies on predictive ability and bias of genomic prediction for boar taint compounds using a single-line and multi-line populations. In general, SNP weighting strategies did not result in better predictive ability for androstenone. On the other hand, considering skatole and indole better predictive ability were archived when using weights based on gene networks. Due the slightly improvement in prediction accuracy and the increase in the number of analyses steps required, the weighting methods may not be advantageous. In addition, we verified that multi-line populations improve the prediction for androstenone, while for skatole and indole this was not observed. Finally, we performed the identification of QTLs and genes associated with boar taint compounds using a weighted single-step genome-wide association study. We used a gene network approach to improve the identification of candidate genes. In summary, we identified the HSD17B2 gene that was previously describe as linked to boar taint appearance. New candidate genes with potential to explain boar taint phenotypes were find: CRHBP, CTDSP2, CDK4, CYP27B1 e SDR4E1. These genes were mainly involved to biosynthesis, releasing and response to steroid hormones and intestinal absorption. Keywords: Androstenone. Candidate gene. Genome-wide selection. Causal relationship. Indole. Skatole.