Overview of the use of genomic data in animal breeding / Panorama do uso de dados genômicos no melhoramento genético animal

ABSTRACT: The use of molecular information in breeding programs contributed to important advances in the improvement of traits of economic interest in livestock production. The advent of single nucleotide polymorphism (SNP) panels applied to genome-wide selection (GWS) and genome-wide association studies (GWAS), along with computational advances (e.g., use of powerful software and robust analyses) allowed a better understanding of the genetic architecture of farm animals and increased the selection efficiency. In this context, the statistic method single-step GBLUP has been frequently used to perform GWS, and more recently GWAS analyses, providing accurate predictions and QTL detection, respectively. Nevertheless, in developing countries, species such as sheep and goats, whose genomic data are more difficult to be obtained, the use of data simulation has been efficient in the study of the major factors involved in the selection process, such as size of training population, density of SNP chips, and genotyping strategies. The effects of these factors are directly associated with the prediction accuracy of genomic breeding values. In this review we showed important aspects of the use of genomics in the genetic improvement of production traits of animals, the main methods currently used for prediction and estimation of molecular marker effects, the importance of data simulation for validation of those methods, as well as the advantages, challenges and limitations of the use of GWS and GWAS in the current scenario of livestock production.

RESUMO: Em programas de melhoramento genético, o uso de informações moleculares garantiu importantes avanços para a melhoria de características de interesse econômico, no âmbito da produção animal. O advento da tecnologia de painéis de SNPs aplicados à seleção genômica ampla (GWS) e associação genômica ampla (GWAS), aliado ao avanço computacional, com o uso de softwares e análises robustas, permitiram melhor compreensão sobre a arquitetura genética dos animais de produção e, consequentemente, maior eficiência na seleção. Nesse contexto, o método estatístico single-step GBLUP tem sido utilizado, frequentemente, na execução da GWS e, mais recentemente, em GWAS, possibilitando predições acuradas e detecção de QTLs, respectivamente. No entanto, em países em desenvolvimento e, em espécies como os ovinos e caprinos, que existe maior dificuldade para a aquisição de dados genômicos, o uso da simulação de dados tem se mostrado eficiente para estudar os principais fatores envolvidos no processo de seleção, como o tamanho da população de treinamento, densidade de chipde SNPs e estratégias de genotipagem, cujos efeitos estão diretamente associados à acurácia da predição de valores genéticos genômicos. Nesta revisão, serão abordados pontos importantes sobre o uso da genômica no melhoramento genético de características produtivas em animais, principais métodos de predição e estimação de efeitos de marcadores moleculares na atualidade, a importância da simulação de dados para a validação desses métodos, bem como as vantagens, os desafios e as limitações no cenário atual da produção animal com o uso da seleção e associação genômica ampla.

Overview of the use of genomic data in animal breeding / Panorama do uso de dados genômicos no melhoramento genético animal

The use of molecular information in breeding programs contributed to important advances in the improvement of traits of economic interest in livestock production. The advent of single nucleotide polymorphism (SNP) panels applied to genome-wide selection (GWS) and genome-wide association studies (GWAS), along with computational advances (e.g., use of powerful software and robust analyses) allowed a better understanding of the genetic architecture of farm animals and increased the selection efficiency. In this context, the statistic method single-step GBLUP has been frequently used to perform GWS, and more recently GWAS analyses, providing accurate predictions and QTL detection, respectively. Nevertheless, in developing countries, species such as sheep and goats, whose genomic data are more difficult to be obtained, the use of data simulation has been efficient in the study of the major factors involved in the selection process, such as size of training population, density of SNP chips, and genotyping strategies. The effects of these factors are directly associated with the prediction accuracy of genomic breeding values. In this review we showed important aspects of the use of genomics in the genetic improvement of production traits of animals, the main methods currently used for prediction and estimation of molecular marker effects, the importance of data simulation for validation of those methods, as well as the advantages, challenges and limitations of the use of GWS and GWAS in the current scenario of livestock production.

Em programas de melhoramento genético, o uso de informações moleculares garantiu importantes avanços para a melhoria de características de interesse econômico, no âmbito da produção animal. O advento da tecnologia de painéis de SNPs aplicados à seleção genômica ampla (GWS) e associação genômica ampla (GWAS), aliado ao avanço computacional, com o uso de softwares e análises robustas, permitiram melhor compreensão sobre a arquitetura genética dos animais de produção e, consequentemente, maior eficiência na seleção. Nesse contexto, o método estatístico single-step GBLUP tem sido utilizado, frequentemente, na execução da GWS e, mais recentemente, em GWAS, possibilitando predições acuradas e detecção de QTLs, respectivamente. No entanto, em países em desenvolvimento e, em espécies como os ovinos e caprinos, que existe maior dificuldade para a aquisição de dados genômicos, o uso da simulação de dados tem se mostrado eficiente para estudar os principais fatores envolvidos no processo de seleção, como o tamanho da população de treinamento, densidade de chipde SNPs e estratégias de genotipagem, cujos efeitos estão diretamente associados à acurácia da predição de valores genéticos genômicos. Nesta revisão, serão abordados pontos importantes sobre o uso da genômica no melhoramento genético de características produtivas em animais, principais métodos de predição e estimação de efeitos de marcadores moleculares na atualidade, a importância da simulação de dados para a validação desses métodos, bem como as vantagens, os desafios e as limitações no cenário atual da produção animal com o uso da seleção e associação genômica ampla.

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.(AU)

QTLs for seedling traits under salinity stress in hexaploid wheat / Caracterização de QTLs por traços de sementes sob o estresse de salinidade em trigo Hexaploid

Soil salinity limits agricultural production and is a major obstacle for increasing crop yield. Common wheat is one of the most important crops with allohexaploid characteristic and a highly complex genome. QTL mapping is a useful way to identify genes for quantitative traits such as salinity tolerance in hexaploid wheat. In the present study, a hydroponic trial was carried out to identify quantitative trait loci (QTLs) associated with salinity tolerance of wheat under 150mM NaCl concentration using a recombinant inbred line population (Xiaoyan 54×Jing 411). Values of wheat seedling traits including maximum root length (MRL), root dry weight (RDW), shoot dry weight (SDW), total dry weight (TDW) and the ratio of TDW of wheat plants between salt stress and control (TDWR) were evaluated or calculated. A total of 19QTLs for five traits were detected through composite interval mapping method by using QTL Cartographer version 2.5 under normal and salt stress conditions. These QTLs distributed on 12 chromosomes explained the percentage of phenotypic variation by individual QTL varying from 7.9% to 19.0%. Among them, 11 and six QTLs were detected under normal and salt stress conditions, respectively and two QTLs were detected for TDWR. Some salt tolerance related loci may be pleiotropic. Chromosome 1A, 3A and 7A may harbor crucial candidate genes associated with wheat salt tolerance. Our results would be helpful for the marker assisted selection to breed wheat varieties with improved salt tolerance.(AU)

A salinidade do solo limita a produção agrícola. O trigo mole é uma das culturas mais importantes com característica allohexaploid e genoma altamente complexo. O mapeamento QTL é uma maneira muito útil de identificar genes para traços quantitativos, como a tolerância à salinidade em trigo hexaplóide. No presente estudo realizou-se um ensaio hidropónico para identificar locos de traços quantitativos (QTLs) associados à tolerância à salinidade do trigo sob concentração de NaCl 150 mM, usando uma população de linhagem consanguíneo recombinante (Xiaoyan 54 × Jing 411). Os valores dos traços de mudas de trigo, incluindo comprimento máximo da raiz (MRL), peso seco da raiz (RDW), ponha o peso seco (SDW), peso seco total (TDW) e a proporção das plantas de trigo TDW entre o estresse salgado e o controle (TDWR), foram avaliados ou calculados. Um total de 19QTLs para cinco traços foram detectados através do método de mapeamento de intervalo composto usando a versão 2.5 do cartógrafo QTL sob condições normais e de estresse salino. Estes QTLs distribuídos em 12 cromossomos explicaram a porcentagem de variação fenotípica por QTL individual variando de 7,9% a 19,0%. Entre eles, foram detectados 11 e 6 QTLs em condições de estresse normal e sal, respectivamente, e dois QTLs foram detectados para TDWR. Cromossoma 1A, 3A e 7A podem conter genes que são candidatos cruciais associados à tolerância ao sal de trigo. Nossos resultados seriam úteis para a seleção assistida por marcadores para produzir variedades de trigo com tolerância salina melhorada.(AU)

QTLs for seedling traits under salinity stress in hexaploid wheat / Caracterização de QTLs por traços de sementes sob o estresse de salinidade em trigo Hexaploid

Soil salinity limits agricultural production and is a major obstacle for increasing crop yield. Common wheat is one of the most important crops with allohexaploid characteristic and a highly complex genome. QTL mapping is a useful way to identify genes for quantitative traits such as salinity tolerance in hexaploid wheat. In the present study, a hydroponic trial was carried out to identify quantitative trait loci (QTLs) associated with salinity tolerance of wheat under 150mM NaCl concentration using a recombinant inbred line population (Xiaoyan 54×Jing 411). Values of wheat seedling traits including maximum root length (MRL), root dry weight (RDW), shoot dry weight (SDW), total dry weight (TDW) and the ratio of TDW of wheat plants between salt stress and control (TDWR) were evaluated or calculated. A total of 19QTLs for five traits were detected through composite interval mapping method by using QTL Cartographer version 2.5 under normal and salt stress conditions. These QTLs distributed on 12 chromosomes explained the percentage of phenotypic variation by individual QTL varying from 7.9% to 19.0%. Among them, 11 and six QTLs were detected under normal and salt stress conditions, respectively and two QTLs were detected for TDWR. Some salt tolerance related loci may be pleiotropic. Chromosome 1A, 3A and 7A may harbor crucial candidate genes associated with wheat salt tolerance. Our results would be helpful for the marker assisted selection to breed wheat varieties with improved salt tolerance.

A salinidade do solo limita a produção agrícola. O trigo mole é uma das culturas mais importantes com característica allohexaploid e genoma altamente complexo. O mapeamento QTL é uma maneira muito útil de identificar genes para traços quantitativos, como a tolerância à salinidade em trigo hexaplóide. No presente estudo realizou-se um ensaio hidropónico para identificar locos de traços quantitativos (QTLs) associados à tolerância à salinidade do trigo sob concentração de NaCl 150 mM, usando uma população de linhagem consanguíneo recombinante (Xiaoyan 54 × Jing 411). Os valores dos traços de mudas de trigo, incluindo comprimento máximo da raiz (MRL), peso seco da raiz (RDW), ponha o peso seco (SDW), peso seco total (TDW) e a proporção das plantas de trigo TDW entre o estresse salgado e o controle (TDWR), foram avaliados ou calculados. Um total de 19QTLs para cinco traços foram detectados através do método de mapeamento de intervalo composto usando a versão 2.5 do cartógrafo QTL sob condições normais e de estresse salino. Estes QTLs distribuídos em 12 cromossomos explicaram a porcentagem de variação fenotípica por QTL individual variando de 7,9% a 19,0%. Entre eles, foram detectados 11 e 6 QTLs em condições de estresse normal e sal, respectivamente, e dois QTLs foram detectados para TDWR. Cromossoma 1A, 3A e 7A podem conter genes que são candidatos cruciais associados à tolerância ao sal de trigo. Nossos resultados seriam úteis para a seleção assistida por marcadores para produzir variedades de trigo com tolerância salina melhorada.

Genome-wide association study using the Bayes C method for important traits in dairy yield in Colombian Holstein Cattle / Associação genômica usando o método Bayes C para características importantes na produção leiteira em Gado Holandês na Colômbia

The objective of this study was to determine the genome association between markers of bovine LD BeadChip with dairy important traits. Information of breeding program of the Universidad Nacional de Colombia was used. BLUP-EBVs were used for dairy yield (DY), fat percentage (FP), proteinpercentage (PP) and somatic cell score (SCS). 150 animals were selected for blood or semen DNA extraction and genotyping with BovineLD BeadChip (Illumina). Autosomal information was retained and the editing information was performed using Plink v1.07 software. The effects of SNPs were determined by Bayes C with GS3 software. The minor allele frequency for most of the markers on the bead chip was high, which increases the probability of finding important loci segregating in the population. Estimations of fraction markers with an effect were close to zero in almost all cases. The most important markers were mapped by trait using ENSEMBL. A total of 6,510 autosomal SNPs were retained, out of which only a proportion with effect was taken from the mixed function of Bayes C. Important genes as ANKS1B, CLCN1, NMBR and CTSD, were found for each trait for AL, FP, PP and SCS respectively. Finally, Bayes C estimation allowed to identify specific SNPs possibly associated with QTLs.(AU)

O Objetivo deste estudo foi determinar a associação genômica entre os marcadores do chip bovino LD (LD BeadChip) com características importantes na produção de leite. Foram utilizadas informações do programa de melhoramento genético da Universidade Nacional de Colômbia. BLUPEBVs foram utilizados para a produção de leite (DY), porcentagem de gordura (FP), porcentagem deproteína (PP) e escore de células somáticas (SCS). 150 animais foram selecionados para extração de DNAdo sangue ou sêmen e genotipados com o chip BovineLD (Illumina). A informação autossômica foi mantidae a edição da informação foi executada usando o programa Plink v1.07. Os efeitos dos SNPs foram determinados por Bayes C com o programa GS3. A frequência do alelo menor para a maioria dos marcadores no chip foi alta, o que aumenta a probabilidade de encontrar locos importantes segregando na população. As estimativas da fração de marcadores, com efeito, foram próximas de zero em quase todas as situações. Os marcadores mais importantes foram mapeados com ENSEMBL. Um total de 6510 SNPs autossômicos foram preservados, dos quais apenas uma proporção foi tomada com efeito a partir da função mista de Bayes C. Para cada característica foram encontrados genes importantes, como ANKS1B, CLCN1, NMBR e CTSD, para AL, FP, PP e SCS, respectivamente. Finalmente, a estimativa de Bayes-C permitiu aidentificação de SNPs com possível associação com QTLs.(AU)

Genome-wide association study using the Bayes C method for important traits in dairy yield in Colombian Holstein Cattle / Associação genômica usando o método Bayes C para características importantes na produção leiteira em Gado Holandês na Colômbia

The objective of this study was to determine the genome association between markers of bovine LD BeadChip with dairy important traits. Information of breeding program of the Universidad Nacional de Colombia was used. BLUP-EBVs were used for dairy yield (DY), fat percentage (FP), proteinpercentage (PP) and somatic cell score (SCS). 150 animals were selected for blood or semen DNA extraction and genotyping with BovineLD BeadChip (Illumina). Autosomal information was retained and the editing information was performed using Plink v1.07 software. The effects of SNPs were determined by Bayes C with GS3 software. The minor allele frequency for most of the markers on the bead chip was high, which increases the probability of finding important loci segregating in the population. Estimations of fraction markers with an effect were close to zero in almost all cases. The most important markers were mapped by trait using ENSEMBL. A total of 6,510 autosomal SNPs were retained, out of which only a proportion with effect was taken from the mixed function of Bayes C. Important genes as ANKS1B, CLCN1, NMBR and CTSD, were found for each trait for AL, FP, PP and SCS respectively. Finally, Bayes C estimation allowed to identify specific SNPs possibly associated with QTLs.

O Objetivo deste estudo foi determinar a associação genômica entre os marcadores do chip bovino LD (LD BeadChip) com características importantes na produção de leite. Foram utilizadas informações do programa de melhoramento genético da Universidade Nacional de Colômbia. BLUPEBVs foram utilizados para a produção de leite (DY), porcentagem de gordura (FP), porcentagem deproteína (PP) e escore de células somáticas (SCS). 150 animais foram selecionados para extração de DNAdo sangue ou sêmen e genotipados com o chip BovineLD (Illumina). A informação autossômica foi mantidae a edição da informação foi executada usando o programa Plink v1.07. Os efeitos dos SNPs foram determinados por Bayes C com o programa GS3. A frequência do alelo menor para a maioria dos marcadores no chip foi alta, o que aumenta a probabilidade de encontrar locos importantes segregando na população. As estimativas da fração de marcadores, com efeito, foram próximas de zero em quase todas as situações. Os marcadores mais importantes foram mapeados com ENSEMBL. Um total de 6510 SNPs autossômicos foram preservados, dos quais apenas uma proporção foi tomada com efeito a partir da função mista de Bayes C. Para cada característica foram encontrados genes importantes, como ANKS1B, CLCN1, NMBR e CTSD, para AL, FP, PP e SCS, respectivamente. Finalmente, a estimativa de Bayes-C permitiu aidentificação de SNPs com possível associação com QTLs.

Quantitative genetics theory for genomic selection and efficiency of genotypic value prediction in open-pollinated populations

Quantitative genetics theory for genomic selection has mainly focused on additive effects. This study presents quantitative genetics theory applied to genomic selection aiming to prove that prediction of genotypic value based on thousands of single nucleotide polymorphisms (SNPs) depends on linkage disequilibrium (LD) between markers and QTLs, assuming dominance and epistasis. Based on simulated data, we provided information on dominance and genotypic value prediction accuracy, assuming mass selection in an open-pollinated population, all quantitative trait loci (QTLs) of lower effect, and reduced sample size. We show that the predictor of dominance value is proportional to the square of the LD value and to the dominance deviation for each QTL that is in LD with each marker. The weighted (by the SNP frequencies) dominance value predictor has greater accuracy than the unweighted predictor. The linear × linear, linear × quadratic, quadratic × linear, and quadratic × quadratic SNP effects are proportional to the corresponding linear combinations of epistatic effects for QTLs and the LD values. LD between two markers with a common QTL causes a bias in the prediction of epistatic values. Compared to phenotypic selection, the efficiency of genomic selection for genotypic value prediction increases as trait heritability decreases. The degree of dominance did not affect the genotypic value prediction accuracy and the approach to maximum accuracy is asymptotic with increases in SNP density. The decrease in the sample size from 500 to 200 did not markedly reduce the genotypic value prediction accuracy.

Quantitative genetics theory for genomic selection and efficiency of genotypic value prediction in open-pollinated populations

Quantitative genetics theory for genomic selection has mainly focused on additive effects. This study presents quantitative genetics theory applied to genomic selection aiming to prove that prediction of genotypic value based on thousands of single nucleotide polymorphisms (SNPs) depends on linkage disequilibrium (LD) between markers and QTLs, assuming dominance and epistasis. Based on simulated data, we provided information on dominance and genotypic value prediction accuracy, assuming mass selection in an open-pollinated population, all quantitative trait loci (QTLs) of lower effect, and reduced sample size. We show that the predictor of dominance value is proportional to the square of the LD value and to the dominance deviation for each QTL that is in LD with each marker. The weighted (by the SNP frequencies) dominance value predictor has greater accuracy than the unweighted predictor. The linear × linear, linear × quadratic, quadratic × linear, and quadratic × quadratic SNP effects are proportional to the corresponding linear combinations of epistatic effects for QTLs and the LD values. LD between two markers with a common QTL causes a bias in the prediction of epistatic values. Compared to phenotypic selection, the efficiency of genomic selection for genotypic value prediction increases as trait heritability decreases. The degree of dominance did not affect the genotypic value prediction accuracy and the approach to maximum accuracy is asymptotic with increases in SNP density. The decrease in the sample size from 500 to 200 did not markedly reduce the genotypic value prediction accuracy.(AU)

Genome association study through nonlinear mixed models revealed new candidate genes for pig growth curves

Genome association analyses have been successful in identifying quantitative trait loci (QTLs) for pig body weights measured at a single age. However, when considering the whole weight trajectories over time in the context of genome association analyses, it is important to look at the markers that affect growth curve parameters. The easiest way to consider them is via the two-step method, in which the growth curve parameters and marker effects are estimated separately, thereby resulting in a reduction of the statistical power and the precision of estimates. One efficient solution is to adopt nonlinear mixed models (NMM), which enables a joint modeling of the individual growth curves and marker effects. Our aim was to propose a genome association analysis for growth curves in pigs based on NMM as well as to compare it with the traditional two-step method. In addition, we also aimed to identify the nearest candidate genes related to significant SNP (single nucleotide polymorphism) markers. The NMM presented a higher number of significant SNPs for adult weight (A) and maturity rate (K), and provided a direct way to test SNP significance simultaneously for both the A and K parameters. Furthermore, all significant SNPs from the two-step method were also reported in the NMM analysis. The ontology of the three candidate genes (SH3BGRL2, MAPK14, and MYL9) derived from significant SNPs (simultaneously affecting A and K) allows us to make inferences with regards to their contribution to the pig growth process in the population studied.

Genome association study through nonlinear mixed models revealed new candidate genes for pig growth curves

Genome association analyses have been successful in identifying quantitative trait loci (QTLs) for pig body weights measured at a single age. However, when considering the whole weight trajectories over time in the context of genome association analyses, it is important to look at the markers that affect growth curve parameters. The easiest way to consider them is via the two-step method, in which the growth curve parameters and marker effects are estimated separately, thereby resulting in a reduction of the statistical power and the precision of estimates. One efficient solution is to adopt nonlinear mixed models (NMM), which enables a joint modeling of the individual growth curves and marker effects. Our aim was to propose a genome association analysis for growth curves in pigs based on NMM as well as to compare it with the traditional two-step method. In addition, we also aimed to identify the nearest candidate genes related to significant SNP (single nucleotide polymorphism) markers. The NMM presented a higher number of significant SNPs for adult weight (A) and maturity rate (K), and provided a direct way to test SNP significance simultaneously for both the A and K parameters. Furthermore, all significant SNPs from the two-step method were also reported in the NMM analysis. The ontology of the three candidate genes (SH3BGRL2, MAPK14, and MYL9) derived from significant SNPs (simultaneously affecting A and K) allows us to make inferences with regards to their contribution to the pig growth process in the population studied.(AU)

Análises genômicas para associação de SNPs com resistência à Aeromonas hydrophila e rendimento de filé no tambaqui (Colossoma macropomum)

Tambaqui (Colossoma macropomum) é o peixe neotropical mais importante da aquicultura de água doce na América do Sul. Surtos da doença causados pela bactéria Aeromonas hydrophila resultam em perdas significativas para a produção de tambaqui. Além disso, o aumento da produção dessa espécie em aquicultura é limitado pela falta de programas de melhoramento genético, especialmente para características de resistência a doenças e rendimento de filé. O objetivo desse estudo foi caracterizar o desempenho zootécnico de famílias de tambaqui C. macropomum para resistência à Aeromonas hydrophila e características de crescimento/rendimento de filé, por meio de genética quantitativa e associação genômica de SNPs (single nucleotide polymorphisms). Primeiramente, os resultados dos parâmetros genéticos para resistência a A. hydrophila e a correlação genética com o ganho de peso diário em juvenis de tambaqui demonstraram que é possível incluir essas características em programas de melhoramento seletivo. Em seguida, para avançar com as análises genômicas, foi desenvolvido um denso SNP array para o tambaqui, com a caracterização de 30 mil SNPs candidatos para a plataforma de genotipagem Axiom/ThermoFischer (30K SerraSNP array). Após essa etapa, foi investigado a arquitetura genética da resistência a A. hydrophila em tambaqui por um estudo de associação do genoma (GWAS) para identificar loci de características quantitativas (QTLs) e genes putativos associados a esta característica em tambaqui. Um denso mapa de ligação foi desenvolvido para o tambaqui usando os dados do genótipo; resultando em 17.374 SNPs distribuídos em 27 grupos de ligação. A análise GWAS identificou vários QTLs candidatos associados à resistência de A. hydrophila, distribuídos em seis grupos de ligação. Vários genes candidatos relacionados à resposta imune foram localizados próximos aos QTLs putativos. Em relação as características de crescimento/rendimento, as análises de genética quantitativa revelaram moderada-alta herdabilidades, o que demonstra que estão sob moderado controle genético e devem responder à seleção genética. Uma abordagem de deep learning e sistema de visão computacional (CVS) também foi validada para obter as medidas morfométricas do tambaqui em larga escala e de forma automatizada. As análises de GWAS para essas características confirmaram que estão sob controle poligênico, com a identificação de apenas alguns QTLs com significância estatística. Em conclusão, ambas as características avaliadas neste estudo, resistência a A. hydrophila e características de crescimento e rendimento corpóreo, demonstram estar sob controle genético e influenciadas por várias regiões genômicas, e podem ser inseridas em programas de melhoramento para tambaqui.

Tambaqui (Colossoma macropomum) is the most important neotropical fish in freshwater aquaculture in South America. Disease outbreaks caused by the bacteria Aeromonas hydrophila result in significant losses for tambaqui production. Furthermore, the increase in production of this species in aquaculture is limited by the lack of genetic improvement programs, especially for disease resistance traits and fillet yield. The objective of this study was to characterize the performance of tambaqui C. macropomum families for resistance to Aeromonas hydrophila and fillet growth/yield traits, through quantitative genetics and genomic association of SNPs (single nucleotide polymorphisms). First, the results of genetic parameters for resistance to A. hydrophila and the genetic correlation with daily weight gain in tambaqui juveniles demonstrated that it is possible to include these traits in selective breeding programs. Then, to proceed with genomic analyses, a dense SNP array for tambaqui was developed, with the characterization of 30,000 candidate SNPs for the Axiom/ThermoFischer genotyping platform (30K SerraSNP array). After this, the genetic architecture of resistance to A. hydrophila in tambaqui was investigated by a genome-wide association study (GWAS) to identify quantitative trait loci (QTLs) and putative genes associated with this trait in tambaqui. A dense linkage map was developed for tambaqui using genotype data; resulting in 17,374 SNPs distributed in 27 linkage groups. GWAS analysis identified several candidate QTLs associated with A. hydrophila resistance, distributed into six linkage groups. Several candidate genes related to the immune response were located close to the putative QTLs. Regarding growth/yield traits, quantitative genetic analysis revealed moderatehigh heritability, which demonstrates that they are under moderate genetic control and must respond to genetic selection. A deep learning approach and computer vision system (CVS) was also validated to obtain the morphometric measurements of tambaqui on a large scale and in an automated way. GWAS analyzes for these traits confirmed that they are under polygenic control, with the identification of only a few QTLs with statistical significance. In conclusion, both traits evaluated in this study, resistance to A. hydrophila and growth and body yield traits, demonstrate to be under genetic control and influenced by several genomic regions, and can be included in improvement programs for tambaqui.

PREDIÇÃO GENÔMICA PONDERADA POR REGIÕES GENÔMICAS CANDIDATAS PARA CARACTERÍSTICAS DE IMPORTÂNCIA ECONÔMICA EM BOVINOS NELORE

Este estudo teve como objetivo avaliar o impacto na acurácia e viés das predições genômicas por incluir e ponderar diferencialmente regiões genômicas candidatas para as características de produção em bovinos da raça Nelore. Foram utilizados dados de 244.254 animais nascidos entre 1977 e 2016, com aproximadamente 176.000 registros fenotípicos para as características peso ajustado aos 450 dias de idade (P450), área de olho de lombo (AOL), marmoreio (MARB) espessura da gordura subcutânea (EGS) e espessura da gordura da garupa (EGP8). 7.769 animais foram genotipados com o painel de baixa densidade (30k - CLARIFIDE® Nelore 3.0) e os genótipos foram imputados a um painel contendo 735.044 marcadores. Para todas as características um modelo animal linear foi considerado para obter as estimativas de parâmetros genéticos e executar o GWAS. Foram incluídos no modelo os efeitos genéticos aditivos diretos e residuais como efeitos aleatórios, os efeitos fixos para as características de carcaça foram grupos contemporâneos (GC) e idade do animal (covariável) e para característica de crescimento os grupos contemporâneos (GC). Para todas as características, um procedimento BLUP genômico de etapa única (ssGBLUP) foi realizado, sendo que a matriz G foi construída usando diferentes combinações de SNPs e pesos. Para avaliar a capacidade de predição do GEBV dos modelos para cada conjunto de dados, a mesma foi mensurada por meio da correlação de Pearson entre o valor genético genômico estimado (GEBV) e o fenótipo ajustado para efeitos fixos (Yc), dividido pela raiz quadrada da herdabilidade da característica avaliada. O viés das predições genômicas foi avaliado utilizando o coeficiente de regressão do fenótipo ajustado (Yc) sobre o GEBV. Verificou-se um maior efeito sobre o viés das predições ao incluir QTLs e ponderações no ssGLBUP do que nas habilidades de predição. Maiores ganhos na habilidade preditiva foram obtidos para P450 ao ponderar a matriz G e incluir QTLs provenientes de ssGWAS. A incorporação de QTLs candidatos previamente relatados na literatura e ponderá-los de forma diferenciada, pode apresentar um sucesso limitado para características de carcaça. Para características com pouca informação e baixa herdabilidade, as estimativas de habilidade preditiva foram baixas, visto que possivelmente o número de observações limita os benefícios da incorporação da informação genômica.

This study aimed to evaluate the impact on the accuracy and bias of genomic predictions by including and differentiately weighting candidate genomic regions for production characteristics in Nelore cattle. We used 244,254 animals born between 1977 and 2016. The data comprise approximately 176,000 phenotypic records for adjusted weight at 450 (P450) days of age, loin eye area (AOL), marbling (MARB) subcutaneous fat thickness (EGS) and croup fat thickness (EGP8). A total of 7,769 animals were genotyped with the low density panel (30k). All genotypes were imputed to a panel containing 735,044 markers. For all traits a linear animal model was considered to obtain the genetic parameter estimates and run GWAS. Direct and residual additive genetic effects were included in the model as random effects, the fixed effects for carcass traits were contemporary groups (CG) and animal age (covariate) and for growth trait the contemporary groups (CG). For all characteristics a single-step genomic BLUP procedure (ssGBLUP) was used. The G Matrix was constructed using different combinations of SNPs and weights. To evaluate the GEBV prediction capability of the models for each data set, it was measured by Pearson's correlation between the estimated genetic genetic value (GEBV) and the fixed effect adjusted phenotype (Yc), divided by the square root of the heritability of the evaluated characteristic. The inflation of the genomic predictions was evaluated using the regression coefficient of the adjusted phenotype (Yc) on GEBV. There was a greater effect on bias of predictions by including candidate QTLs and weights in ssGLBUP than in prediction abilities. More gains in predictive ability were obtained for P450 by weighting the G matrix and including candidate QTLs from ssGWAS. Incorporating QTLs previously reported in the literature and weighting them differently may have limited success for carcass traits. For traits with little information and low heritability, predictive ability estimates were low, as possibly the number of observations limits the benefits of incorporating genomic information.

Identificação e uso de CNV para seleção genômica em ovinos da raça Santa Inês

Variações no número de cópias (CNV) são variações estruturais no DNA de no mínimo 1 kb e que diferem do genoma referência quanto ao número de cópias. O crescente número de estudos com CNV reportam sua associação com doenças em humanos e diversas características quantitativas em animais domésticos. Entretanto, não foram reportados estudos de associação genômica ampla baseados em CNV em ovinos Santa Inês, uma das raças mais importantes do Brasil e, ademais, informações sobre a inclusão de efeitos de CNV em modelos de predição de GEBV são escassos. Assim, os objetivos do presente estudo foram 1) identificar e quantificar CNVs em uma população de ovinos da raça Santa Inês, 2) verificar possíveis associações entre CNVs e características de interesse, 3) identificar possível sobreposição de segmentos de CNVs significativamente associados às características de interesse a QTLs e genes previamente conhecidos, e 4) avaliar possíveis ganhos em acurácia das predições dos GEBVs combinando informações de SNP e CNV (GBa + CNV) comparadas às predições dos GEBVs utilizando exclusivamente informações SNP (GBa). Dessa forma, foram utilizadas informações genômicas de 491 ovinos da raça Santa Inês (OvineSNP50 BeadChip). A identificação de CNVs foi realizada por meio do software PennCNV. As informações de CNVs foram utilizadas para a construção de uma matriz de relacionamento genômico fundamentada em CNVs que foi incorporada ao modelo misto utilizado para a predição dos GEBVs adicionalmente às tradicionais informações de SNP. Na sequência, a acurácia foi estimada pela correlação de Pearson entre os GEBVs preditos para a população de treinamento e o fenótipo corrigido para efeitos fixos na população de validação. Os GEBVs preditos foram utilizados nas análises de associação genômica ampla fundamentadas em CNVs, por meio do programa CNVRanger, para identificar regiões no genoma associadas às características de desempenho, eficiência e carcaça. Foram identificados 1.167 segmentos de CNV, entre eles segmentos significantes (p-valor<0,05) associados às características: rendimento de carcaça e consumo alimentar residual, adicionalmente, outros 5 segmentos de CNV foram considerados relevantes e indicados para estudos futuros. Os segmentos significativos foram sobrepostos a 4 QTLs e 8 genes, entre eles os genes SPAST, TGFA e ADGRL3 com funções relacionadas a diferenciação celular e metabolismo energético. A inclusão de CNV nas análises genéticas permitiram capturar maior variância genética, entretanto a proporção da variância genética capturada por CNVs varia entre as características e não necessariamente incrementam a acurácia de predição. Entre os modelos considerados, não houve incremento na habilidade preditiva ao inserir informações de CNV. Portanto, os resultados reportados no presente estudo indicam que a inclusão de informações de CNV em modelos para predição de GEBV não deve ser prioridade.

Copy number variations (CNV) are structural variations in DNA of at least 1 kb that differ from the reference genome in copy number. The growing number of studies with CNV report its association with diseases in humans and several quantitative traits in livestock animals. However, CNV-based genome-wide association studies in Santa Inês sheep, one of the most important breeds in Brazil, have not been reported and, furthermore, information on the inclusion of CNV effects in GEBV prediction models is rare. Thus, the aims of the present study were 1) to identify and quantify CNVs in a population of Santa Inês sheep, 2) to verify possible associations between CNVs and characteristics of interest, 3) to identify possible overlapping of CNV segments significantly associated with the characteristics of interest to previously known QTLs and genes, and 4) to evaluate possible gains in accuracy in GEBV predictions by combining SNP and CNV information (GBa + CNV) compared to GEBV predictions using exclusively SNP information (GBa). Thus, genomic information from 491 Santa Inês sheep (OvineSNP50 BeadChip) was used. The identification of CNVs was performed using the PennCNV software. The CNV information was used to build a genomic relationship matrix based on CNVs, which was incorporated into the mixed model used for the prediction of GEBVs in addition to the traditional SNP information. Next, accuracy was estimated by Pearson's correlation between predicted GEBVs for training population and phenotype corrected for fixed effects in the validation population. Predicted GEBVs were used in CNV-based genome-wide association analyses, through the CNVRanger program, to identify regions in the genome associated with performance, efficiency and carcass traits. 1,167 CNV segments were identified, including significant segments (p-value<0.05) associated with traits: carcass yield and residual feed intake, in addition, another 5 CNV segments were considered relevant and indicated for future studies. The significant segments overlapped 4 QTLs and 8 genes, including the SPAST, TGFA and ADGRL3 genes with functions related to cell differentiation and energy metabolism. The inclusion of CNV in the genetic analysis allowed capturing greater genetic variance, however the proportion of genetic variance captured by CNVs varies among characteristics and does not necessarily increase the prediction accuracy. Among the models considered, there was no increase in predictive ability when entering CNV information. Therefore, the results reported in the present study indicate that the inclusion of CNV information in models to predict GEBV should not be a priority.

MARCADORES MOLECULARES NA IDENTIFICAÇÃO DE CORRIDAS DE HOMOZIGOSE E ENDOGAMIA NAS RAÇAS CURRALEIRO PÉ-DURO E PANTANEIRO

As corridas de homozigose (ROHs - runs of homozygosity) podem ser utilizadas para estimar parâmetros populacionais mais precisos e regiões genômicas sob seleção em diferentes espécies e raças. Sendo assim, essa abordagem pode ser usada em raças localmente adaptadas para identificar genes de interesse econômico e estimar autozigosidade possibilitando a otimização das estratégias de seleção e acasalamento e mitigando o seu risco de extinção. O objetivo do presente trabalho é identificar e caracterizar as ROHs presentes nas raças localmente adaptadas Curraleiro Pé-Duro (CPD) e Pantaneiro (PANT) com o intuito de identificar genes sob seleção e estimar seu coeficiente de endogamia (FROH). As ROHs foram identificadas em 126 e 35 animais das raças CPD e PANT, respectivamente, com dados de 54.000 marcadores SNP. Para a identificação das ROHs foram utilizados os seguintes critérios: a) janela deslizante de 50 SNPs, b) número mínimo de 100 SNPs consecutivos em cada ROH, c) comprimento mínimo de 1 Mb, c) intervalo máximo entre SNPs homozigotos consecutivos de 0,5 Mb, d) máximo de cinco SNPs com genótipos ausentes, Foram encontradas 1083 e 140 ROHs nas raças CPD e PANT, respectivamente, e que o padrão predominante de ROHs nas raças estudadas são as de tamanho médio a longo o que demonstra eventos recentes de autozigosidade. Foram identificadas seis ilhas de homozigose em ambas as raças nos cromossomo autossômicos 6, 8, 10, 13, 14 e 17. Genes e QTLs previamente relacionados com características de interesse estão localizados nos segmentos encontrados, podendo destacar QTLs associados à caracteristicas de produção e qualidade de leite (57%), como os envolvidos no teor de proteínas como as caseínas e genes associados ao sistema imune, como os genes da família de beta defensinas. O coeficente de endogamia estimado demonstrou que ambas as raças apresentam baixa autozigosidade (2-4%), o que difere do esperado para populações de base genética reduzida como as estudadas. O presente trabalho identificou seis ilhas de homozigose em comum nas raças Curraleiro Pé-Duro e Pantaneiro, localizadas em regiões genômicas previamente relatadas como associadas a características de interesse econômico, como resistência a doenças e teor de proteína no leite e que podem ser consideradas regiões candidatas à seleção. O coeficiente de endogamia estimado com base nas ROHs, demonstrou que a endogamia é baixa (2 a 4%) nas populações estudadas, divergindo do esperado em pequenas populações.

The runs of homozygosity (ROHs) can be used to estimate more accurate population parameters and genomic regions under selection in different species and breeds. Therefore, this approach can be used in locally adapted breeds to identify genes of economic interest and estimate autozygosity, enabling the optimization of selection and mating strategies and mitigating their extinction risk. The objective of this work is to identify and characterize the ROHs present in locally adapted breeds Curraleiro Pé-Duro (CPD) and Pantaneiro (PANT) in order to identify genes under selection and estimate their inbreeding coefficient (FROH). The ROHs were identified in 126 and 35 animals of the CPD and PANT breeds, respectively, with data from 54,000 SNP markers. For the identification of ROHs, the following criteria were used: a) sliding window of 50 SNPs, b) minimum number of 100 consecutive SNPs in each ROH, c) minimum length of 1 Mb, c) maximum interval between homozygous consecutive SNPs of 0, 5 Mb, d) maximum of five SNPs with absent genotypes, 1083 and 140 ROHs were found in the CPD and PANT breeds, respectively, and that the predominant pattern of ROHs in the studied breeds are those of medium to long size, which demonstrates recent events of autozygosity. Six islands of homozygosity were identified in both breeds on autosomal chromosomes 6, 8, 10, 13, 14 and 17. Genes and QTLs previously related to traits of interest are located in the segments found, highlighting QTLs associated with production and quality traits of milk (57%), as those involved in the content of proteins such as caseins and genes associated with the immune system, such as the genes of the beta defensin family. The estimated inbreeding coefficient showed that both breeds have low autozygosity (2-4%), which differs from what is expected for populations with a reduced genetic base such as those studied. The present work identified six homozygous islands in common in the Curraleiro Pé-Duro and Pantaneiro breeds, located in genomic regions previously reported to be associated with characteristics of economic interest, such as disease resistance and milk protein content, which can be considered candidate regions the selection. The inbreeding coefficient estimated based on the ROHs showed that inbreeding is low (2 to 4%) in the studied populations, diverging from what was expected in small populations.

Quantitative trait locus mapping associated with earliness and fruit weight in tomato

The flowering time is regarded as an important factor that affects yield in various crops. In order to understand how the molecular basis controlling main components of earliness in tomato (Solanum lycopersicum L.), and to deduce whether the correlation between fruit weight, days to flowering and seed weight, is caused by pleiotropic effects or genetic linkage, a QTLs analysis was carried out using an F2 interspecific population derived from the cross of S. lycopersicum and S. pimpinellifolium. The analysis revealed that most of the components related to earliness were independent due to the absence of phenotypic correlation and lack of co-localization of their QTLs. QTLs affecting the flowering time showed considerable variation over time in values of explained phenotypic variation and average effects, which suggested dominance becomes more evident over time. The path analysis showed that traits such as days to flowering, seed weight, and length of the first leaf had a significant effect on the expression of fruit weight, confirming that their correlations were due to linkage. This result was also confirmed in two genomic regions located on chromosomes 1 and 4, where despite showing high co-localization of QTLs associated to days to flowering, seed weight and fruit weight, the presence and absence of epistasis in dfft1.1 × dftt4.1 and fw1.1 × fw4.1, suggested that the linkage was the main cause of the co-localization.

Quantitative trait locus mapping associated with earliness and fruit weight in tomato

The flowering time is regarded as an important factor that affects yield in various crops. In order to understand how the molecular basis controlling main components of earliness in tomato (Solanum lycopersicum L.), and to deduce whether the correlation between fruit weight, days to flowering and seed weight, is caused by pleiotropic effects or genetic linkage, a QTLs analysis was carried out using an F2 interspecific population derived from the cross of S. lycopersicum and S. pimpinellifolium. The analysis revealed that most of the components related to earliness were independent due to the absence of phenotypic correlation and lack of co-localization of their QTLs. QTLs affecting the flowering time showed considerable variation over time in values of explained phenotypic variation and average effects, which suggested dominance becomes more evident over time. The path analysis showed that traits such as days to flowering, seed weight, and length of the first leaf had a significant effect on the expression of fruit weight, confirming that their correlations were due to linkage. This result was also confirmed in two genomic regions located on chromosomes 1 and 4, where despite showing high co-localization of QTLs associated to days to flowering, seed weight and fruit weight, the presence and absence of epistasis in dfft1.1 × dftt4.1 and fw1.1 × fw4.1, suggested that the linkage was the main cause of the co-localization.(AU)

Quantitative genetics theory for genomic selection and efficiency of breeding value prediction in open-pollinated populations

To date, the quantitative genetics theory for genomic selection has focused mainly on the relationship between marker and additive variances assuming one marker and one quantitative trait locus (QTL). This study extends the quantitative genetics theory to genomic selection in order to prove that prediction of breeding values based on thousands of single nucleotide polymorphisms (SNPs) depends on linkage disequilibrium (LD) between markers and QTLs, assuming dominance. We also assessed the efficiency of genomic selection in relation to phenotypic selection, assuming mass selection in an open-pollinated population, all QTLs of lower effect, and reduced sample size, based on simulated data. We show that the average effect of a SNP substitution is proportional to LD measure and to average effect of a gene substitution for each QTL that is in LD with the marker. Weighted (by SNP frequencies) and unweighted breeding value predictors have the same accuracy. Efficiency of genomic selection in relation to phenotypic selection is inversely proportional to heritability. Accuracy of breeding value prediction is not affected by the dominance degree and the method of analysis, however, it is influenced by LD extent and magnitude of additive variance. The increase in the number of markers asymptotically improved accuracy of breeding value prediction. The decrease in the sample size from 500 to 200 did not reduce considerably accuracy of breeding value prediction.

Quantitative genetics theory for genomic selection and efficiency of breeding value prediction in open-pollinated populations

To date, the quantitative genetics theory for genomic selection has focused mainly on the relationship between marker and additive variances assuming one marker and one quantitative trait locus (QTL). This study extends the quantitative genetics theory to genomic selection in order to prove that prediction of breeding values based on thousands of single nucleotide polymorphisms (SNPs) depends on linkage disequilibrium (LD) between markers and QTLs, assuming dominance. We also assessed the efficiency of genomic selection in relation to phenotypic selection, assuming mass selection in an open-pollinated population, all QTLs of lower effect, and reduced sample size, based on simulated data. We show that the average effect of a SNP substitution is proportional to LD measure and to average effect of a gene substitution for each QTL that is in LD with the marker. Weighted (by SNP frequencies) and unweighted breeding value predictors have the same accuracy. Efficiency of genomic selection in relation to phenotypic selection is inversely proportional to heritability. Accuracy of breeding value prediction is not affected by the dominance degree and the method of analysis, however, it is influenced by LD extent and magnitude of additive variance. The increase in the number of markers asymptotically improved accuracy of breeding value prediction. The decrease in the sample size from 500 to 200 did not reduce considerably accuracy of breeding value prediction.(AU)