Quantitative genetics and genomic tools applied to aquaculture breeding
Carregando...
Data
2021-11-04
Autores
Garcia Neto, Baltasar Fernandes
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A procura por animais mais uniformes frente as adversidades ambientais e a aplicação de ferramentas genômicas são exemplos de novas estratégias para tornar o melhoramento genético em espécies aquícolas mais eficiente. No intuito de avaliar a efetividade destas estratégias, os objetivos do presente trabalho foram: i) estimar os componentes genéticos da uniformidade de peso a despesca (PD) e investigar se o PD e sua uniformidade (PDu) podem afetar a sobrevivência (SOB) em camarões; ii) estimar o desequilíbrio de ligação (LD) em uma população de camarões cultivados usando um novo painel de 50k polimorfismos de nucleotídeos únicos (SNPs) e iii) avaliar a viabilidade da imputação ao nível de sequência em tilápia do Nilo e os impactos do tamanho e origem da população de referência na acurácia. Para i), foram utilizados 149.919 registros de PD e 164.023 de SOB de uma larvicultura de camarões mexicana. Os dados foram agrupados em três conjuntos de acordo com o tipo de produção: viveiros escavados com baixa densidade (S1), tanques de concreto com recirculação em alta densidade (S2) e juntando ambos conjuntos de dados (S1+S2). Um modelo linear generalizado hierárquico duplo foi aplicado para estimar os componentes de (co)variância de PD-PDu e um modelo bivariado linear misto foi utilizado para estimar os componentes referentes a PD-SOB. As correlações genéticas (rg) para estas características foram estimadas, além da correlação entre valores genéticos (VG) para PDu-SOB. Para ii), 96 camarões (40 machos e 56 fêmeas) reprodutores originários de uma larvicultura do Equador foram genotipados utilizando um novo painel de 50k SNPs. Posteriormente, o LD foi estimado usando três controles de qualidade distintos com diferentes filtros de frequência de alelo menor: 0,1 (CQ1); 0,05 (CQ2) e 0,01 (CQ3). Para iii), amostras de DNA de 326 tilápias provenientes de três populações de origens distintas (PA, PB e PC) foram extraídas e sequenciadas. Após o controle de qualidade dos dados de sequência foram obtidos 4,6 milhões de SNPs em comum para todas as populações. A imputação foi feita em quatro cenários distintos: dois tamanhos (10 ou 90% dos animais de cada população) e duas origens de referência (apenas duas populações diferentes ou todas as três populações). Os animais de validação tiveram seus genótipos ocultados mantendo somente 50k SNPs para a imputação a nível de sequência. Foi utilizado o software FImpute3 e a acurácia de imputação foi avaliada através da correlação entre o genótipo imputado e observado (r²). No estudo de uniformidade, uma proporção significativa de variância genética foi detectada na variância residual de PD, revelando que existe a possibilidade de selecionar para uniformidade desta característica em camarões (coeficiente de variação genética variando de 17 a 35%). As rg entre PD e SOB foram diferentes em sinal e magnitude com base no sistema de produção, sendo iguais a 0,36, -0,59 e -0,02 para S1, S2 e S1+S2, respectivamente. Os coeficientes de correlação entre o VG de PDu e SOB foram significativos apenas para S1 (-0,32), sugerindo que a seleção para famílias mais uniformes também pode aumentar as taxas de sobrevivência em um ambiente de menor densidade. Para os dados genômicos de camarões, foram obtidos 34.425, 39.091 e 42.789 SNPs após CQ1, CQ2 e CQ3, respectivamente, exibindo alto grau de polimorfismo e validando a aplicação deste painel de SNPs para esta população de camarões cultivados. O LD decaiu rapidamente nos primeiros 30 KB de distância de 0,2 para 0,07 e depois diminuiu para 0,02 para distancias mais longas (>80 KB). Esses resultados sugerem incorporação recente de animais de diferentes populações ou linhagens neste grupo de reprodutores e que a seleção genômica e estudos de associação amplo do genoma são viáveis em camarões usando este painel de SNP como ferramenta. Para a imputação, no geral, o r² por animal mostrou resultados intermediários variando de 0,37 a 0,56 para PA e 0,43 a 0,58 para PB e Pc. Entretanto, os resultados mostraram que foi possível imputar de 50k a aproximadamente 680k com alta acurácia usando dados de sequência de tilápia. Foi observado um aumento de 31,5% para PA e 24,6% para PB e PC no r², quando 90% dos animais da mesma população foram usados como referência em comparação a 10%. Não houve diferenças significativas para r² entre os cenários que usaram 90% dos animais da mesma população e usaram animais das três populações como referência, mostrando que a estratégia de usar informações de outra população para aumentar a população de referência teve pouco efeito na acurácia de imputação.
The search for more uniform animals in the face of environmental adversities and the application of genomic tools are examples of new strategies to make genetic improvement in aquaculture species more efficient. In order to evaluate the effectivity of these strategies, the aims of the present study were: i) to estimate the genetic components of uniformity of harvest weight (HW) and investigate whether the HW and its uniformity (HWv) may affect the survival (SUR) in shrimp; ii) to estimate linkage disequilibrium (LD) in a population of farmed shrimp using a new 50k single-nucleotide polymorphisms (SNPs) panel and; iii) to assess the feasibility of genotype imputation to sequence level in Nile tilapia and the impacts of size and origin of population reference on accuracy of imputation. For i), 149,919 records of HW and 164,023 of SUR were obtained from a Mexican shrimp hatchery. The data were grouped into three sets according to the system of production: excavated ponds with low density (S1), concrete recirculation tanks with high density (S2) and joining both sets of data (S1+S2). A double hierarchical generalized linear model was applied to estimate the (co)variance components of HW-HWv and a bivariate mixed linear model was used to estimate the components referring to HW-SUR. The genetic correlations (rg) for these traits were estimated, in addition to the correlation between the estimated breeding values (EBV) for HWv-SUR. For ii), 96 broodstock shrimp (40 males and 56 females) from a commercial hatchery from Ecuador were genotyped using a new 50k SNPs panel. Then, the LD was estimated using three distinct quality controls with different minor allele frequency filters: 0.1 (QC1); 0.05 (QC2) and 0.01 (QC3). For iii), DNA samples from 326 tilapia from three populations (PA, PB and PC) were extracted and sequenced. After quality control of sequence data, 4.6 million of SNPs were obtained in common for all populations. The imputation was performed in four different scenarios: two sizes (10 or 90% of animals from each population) and two origins of population reference (only two different populations or all three populations). The validation animals had their genotypes masked keeping only 50k SNPs for imputation to the sequence level. The FImpute3 software was used and the imputation accuracy was evaluated through the correlation between the imputed and observed genotypes (r²). In the uniformity study, a significant proportion of genetic variance was detected in the residual variance of HW, revealing that there is possibility to select for uniformity of this trait in shrimp (genetic coefficient of variation ranging between 17 and 35%). The genetic correlations between HW and SUR were different in sign and magnitude based on the production system, being equal to 0.36, -0.59 and -0.02 for S1, S2 and S1+S2, respectively. The correlation coefficients between the EBV of HWv and SUR were significant only for S1 (-0.32), suggesting that selection for more uniform families may also increase survival rates in a lower density environment. For shrimp genomic data, 34,425, 39,091 and 42,789 SNPs were obtained after QC1, QC2 and QC3, respectively, showing high polymorphism and validating the application of this panel of SNPs to this farmed shrimp population. The LD declined rapidly in the first 30 KB of distance from 0.2 to 0.07 and then decreased to 0.02 for longer distances (>80 KB). These results suggest recent incorporation of animals from different populations or strains into this breeding population and that genomic selection and genome-wide association studies are feasible in shrimp using this SNP panel as tool. For the imputation using tilapia sequence data, in general, the r² per animal showed intermediate results ranging from 0.37 to 0.56 for PA and 0.43 to 0.58 for PB and PC. However, the results showed that it was possible to impute from 50k to approximately 680k with high accuracy. An increase in r² of 31.5% for PA and 24.6% for PB and PC was observed, when 90% of animals from the same population were used as reference. There were no significant differences for r² between the scenarios that used 90% of animals from the same population and used animals from the three populations as a reference, showing that the strategy of using information from another population to increase the reference population had minor effect on the accuracy of imputation.
The search for more uniform animals in the face of environmental adversities and the application of genomic tools are examples of new strategies to make genetic improvement in aquaculture species more efficient. In order to evaluate the effectivity of these strategies, the aims of the present study were: i) to estimate the genetic components of uniformity of harvest weight (HW) and investigate whether the HW and its uniformity (HWv) may affect the survival (SUR) in shrimp; ii) to estimate linkage disequilibrium (LD) in a population of farmed shrimp using a new 50k single-nucleotide polymorphisms (SNPs) panel and; iii) to assess the feasibility of genotype imputation to sequence level in Nile tilapia and the impacts of size and origin of population reference on accuracy of imputation. For i), 149,919 records of HW and 164,023 of SUR were obtained from a Mexican shrimp hatchery. The data were grouped into three sets according to the system of production: excavated ponds with low density (S1), concrete recirculation tanks with high density (S2) and joining both sets of data (S1+S2). A double hierarchical generalized linear model was applied to estimate the (co)variance components of HW-HWv and a bivariate mixed linear model was used to estimate the components referring to HW-SUR. The genetic correlations (rg) for these traits were estimated, in addition to the correlation between the estimated breeding values (EBV) for HWv-SUR. For ii), 96 broodstock shrimp (40 males and 56 females) from a commercial hatchery from Ecuador were genotyped using a new 50k SNPs panel. Then, the LD was estimated using three distinct quality controls with different minor allele frequency filters: 0.1 (QC1); 0.05 (QC2) and 0.01 (QC3). For iii), DNA samples from 326 tilapia from three populations (PA, PB and PC) were extracted and sequenced. After quality control of sequence data, 4.6 million of SNPs were obtained in common for all populations. The imputation was performed in four different scenarios: two sizes (10 or 90% of animals from each population) and two origins of population reference (only two different populations or all three populations). The validation animals had their genotypes masked keeping only 50k SNPs for imputation to the sequence level. The FImpute3 software was used and the imputation accuracy was evaluated through the correlation between the imputed and observed genotypes (r²). In the uniformity study, a significant proportion of genetic variance was detected in the residual variance of HW, revealing that there is possibility to select for uniformity of this trait in shrimp (genetic coefficient of variation ranging between 17 and 35%). The genetic correlations between HW and SUR were different in sign and magnitude based on the production system, being equal to 0.36, -0.59 and -0.02 for S1, S2 and S1+S2, respectively. The correlation coefficients between the EBV of HWv and SUR were significant only for S1 (-0.32), suggesting that selection for more uniform families may also increase survival rates in a lower density environment. For shrimp genomic data, 34,425, 39,091 and 42,789 SNPs were obtained after QC1, QC2 and QC3, respectively, showing high polymorphism and validating the application of this panel of SNPs to this farmed shrimp population. The LD declined rapidly in the first 30 KB of distance from 0.2 to 0.07 and then decreased to 0.02 for longer distances (>80 KB). These results suggest recent incorporation of animals from different populations or strains into this breeding population and that genomic selection and genome-wide association studies are feasible in shrimp using this SNP panel as tool. For the imputation using tilapia sequence data, in general, the r² per animal showed intermediate results ranging from 0.37 to 0.56 for PA and 0.43 to 0.58 for PB and PC. However, the results showed that it was possible to impute from 50k to approximately 680k with high accuracy. An increase in r² of 31.5% for PA and 24.6% for PB and PC was observed, when 90% of animals from the same population were used as reference. There were no significant differences for r² between the scenarios that used 90% of animals from the same population and used animals from the three populations as a reference, showing that the strategy of using information from another population to increase the reference population had minor effect on the accuracy of imputation.
Descrição
Palavras-chave
Camarão, Desequilíbrio de ligação, Imputação de genótipos, Tilápia, Uniformidade, Shrimp, Linkage Disequilibrium, Genotype Imputation, Uniformity