Métodos aplicados aos estudos de associação genômica via regiões cromossômicas considerando efeitos aditivos e de dominância
Arquivos
Data
2021-03-01
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Viçosa
Resumo
Os avanços na biologia molecular e as inovações nas tecnologias de sequenciamento e de genotipagem permitiram o desenvolvimento de novos marcadores moleculares favorecendo os estudos de associação genômica ampla (Genome Wide Association Studies - GWAS). A análise via marcas únicas se destaca como o principal procedimento para estudar a associação entre marcas e QTL (Quantitative Trait Loci), porém metodologias que consideraram grupos de marcadores para flanquear regiões genômicas vem elucidando importantes resultados para estudos de associação. Várias abordagens estatísticas veem sendo propostas no âmbito da GWAS, no entanto, estudos comparativos revelam que os métodos bayesianos são superiores em termos do poder em detectar marcadores com associações significativas. Entre os critérios existentes de seleção de regiões se destacam, a seleção pela porcentagem da variância explicada por regiões genômicas (%var), o critério de seleção de tagSNPs (tagSNPs) e a seleção com base na probabilidade a posteriori da associação de regiões genômicas (WPPA - Window Posterior Probability of Association). Para também detectar regiões potencialmente associadas, foi proposto o critério baseado na probabilidade a posteriori do intervalo (Posterior Probability of Interval - 𝑃𝑃 𝑖𝑛𝑡 ) que visa selecionar regiões com base nos marcadores de maiores efeitos estimados via método bayesiano, neste estudo o BayesD𝜋. Além disso, uma metodologia alternativa, denominada mapeamento de herdabilidades regionais (Regional Heritability Mapping - RHM) vem apresentando importantes resultados. Dessa forma, o primeiro capítulo deste trabalho consiste em uma revisão de literatura sobre a GWAS apresentando sua definição e importância no melhoramento genético e abordando detalhes teóricos acerca dos critérios citados acima. Já o capítulo 2 visa propor a medida 𝑃𝑃 𝑖𝑛𝑡 e compará-la às demais abordagens, tagSNP, %var, WPPA conjuntamente ao BayesD𝜋 e metodologia de marcas únicas, quanto a eficiência em selecionar e identificar marcadores ou regiões associados a QTL. Para isso, utilizou-se dados simulados considerando seis cenários diferentes, sendo os SNPs alocados em regiões genômicas não sobrepostas. Os resultados do segundo capítulo indicaram que para características com herança oligogênica, o critério WPPA seguido dos critérios %var e 𝑃𝑃 𝑖𝑛𝑡 se mostraram superiores, apresentando maiores valores de poder de detecção, capturando maiores porcentagens de variância genética e maiores áreas. Para características com herança poligênica, os critérios 𝑃𝑃 𝑖𝑛𝑡 e WPPA foram considerados superiores aos demais. Ademais, o capítulo 3 avalia os critérios, 𝑃𝑃 𝑖𝑛𝑡 e WPPA, que se mostraram superiores no capítulo 2 junto aos métodos de análise via marcas únicas e o RHM. No entanto, a eficiência em termos de poder de detecção e de falsos positivos destes métodos foi avaliada considerando ou não a inclusão dos efeitos de dominância nos modelos estatísticos. Para isso, foram utilizados dados simulados em dezoito cenários com diferentes níveis de herdabilidade, arquitetura genética e grau médio de dominância. Os resultados indicaram que para os efeitos aditivos considerando características com arquitetura genética oligogênica, os critérios WPPA, RHM e 𝑃𝑃 𝑖𝑛𝑡 se mostraram superiores para todos os graus de dominância analisados. Já para características com herança poligênica, os critérios 𝑃𝑃 𝑖𝑛𝑡 e WPPA podem ser considerados superiores aos demais. Considerando apenas os efeitos devido à dominância, os critérios WPPA, RHM, análise via marcas únicas e 𝑃𝑃 𝑖𝑛𝑡 apresentaram resultados relevantes com relação as medidas de eficiência para as características controladas por 3 QTL. Palavras-chave: Regiões Genômicas. Marcadores Moleculares. Métodos Bayesianos. Variância Genética.
Advances in molecular biology and innovations in sequencing and genotyping technologies have allowed the development of new molecular markers favoring genome-wide association studies (GWAS). The single-marker analyses stand out as the central procedure to study the association between markers and quantitative trait loci (QTL). However, methodologies that considered groups of markers to flank genomic regions have elucidated important results for association studies. Several statistical approaches are being proposed within the scope of GWAS. However, comparative analyses reveal that Bayesian methods are superior in terms of the power to detect markers with significant associations. Among the existing criteria for the region selection, the selection by the percentage of variance explained (%var), the selection criteria for tag single nucleotide polymorphisms (tagSNPs), and the selection based on the window posterior probability of association (WPPA). To also detect potentially associated regions, a criterion based on the a posteriori probability of interval (𝑃𝑃 𝑖𝑛𝑡 ) was proposed, aiming to select regions based on the markers of greatest effects estimated via the Bayesian method. In this study, the BayesDπ. An alternative methodology, called regional heritability mapping (RHM), has been shown substantial results. Thus, the first chapter of this work consists of a literature review on GWAS presenting its definition and importance in genetic improvement and addressing theoretical details about the criteria mentioned above. Chapter 2 aims to propose the 𝑃𝑃 𝑖𝑛𝑡 measure and compare it to the other approaches, tagSNP, %var, WPPA together with BayesDπ and single-marker analyses, regarding the efficiency in selecting and identifying markers or regions associated with QTL. For this, simulated data was used considering six different scenarios, with SNPs being allocated in non-overlapping genomic regions. The second chapter results indicated that for traits with oligogenic inheritance, the WPPA criterion followed by the %var and 𝑃𝑃 𝑖𝑛𝑡 criteria were superior, presenting higher values of detection power, capturing higher percentages of genetic variance and larger areas. The criteria 𝑃𝑃 𝑖𝑛𝑡 and WPPA were considered superior to the others. Also, chapter 3 evaluates the criteria, 𝑃𝑃 𝑖𝑛𝑡 and WPPA, which proved to be superior in chapter 2 together with the single- marker analyses and RHM. However, the detection power and the false positives was assessed considering whether (or not) the inclusion of the dominance effects in the statistical models. For that, simulated data were used in eighteen scenarios with different heritability levels, genetic architecture, and degree of dominance. The results indicated that for the additive effects considering traits with oligogenic genetic architecture, the WPPA, RHM and 𝑃𝑃 𝑖𝑛𝑡 criteria were superior for all analyzed degrees of dominance. For traits with polygenic inheritance, the 𝑃𝑃 𝑖𝑛𝑡 and WPPA criteria can be considered superior. Considering only the effects due to dominance, the criteria WPPA, RHM, single-marker analyses and 𝑃𝑃 𝑖𝑛𝑡 presented relevant results regarding the efficiency measures for the traits controlled by 3 QTL. Keywords: Genomic Regions. Molecular Markers. Bayesian Methods. Genetic Variance.
Advances in molecular biology and innovations in sequencing and genotyping technologies have allowed the development of new molecular markers favoring genome-wide association studies (GWAS). The single-marker analyses stand out as the central procedure to study the association between markers and quantitative trait loci (QTL). However, methodologies that considered groups of markers to flank genomic regions have elucidated important results for association studies. Several statistical approaches are being proposed within the scope of GWAS. However, comparative analyses reveal that Bayesian methods are superior in terms of the power to detect markers with significant associations. Among the existing criteria for the region selection, the selection by the percentage of variance explained (%var), the selection criteria for tag single nucleotide polymorphisms (tagSNPs), and the selection based on the window posterior probability of association (WPPA). To also detect potentially associated regions, a criterion based on the a posteriori probability of interval (𝑃𝑃 𝑖𝑛𝑡 ) was proposed, aiming to select regions based on the markers of greatest effects estimated via the Bayesian method. In this study, the BayesDπ. An alternative methodology, called regional heritability mapping (RHM), has been shown substantial results. Thus, the first chapter of this work consists of a literature review on GWAS presenting its definition and importance in genetic improvement and addressing theoretical details about the criteria mentioned above. Chapter 2 aims to propose the 𝑃𝑃 𝑖𝑛𝑡 measure and compare it to the other approaches, tagSNP, %var, WPPA together with BayesDπ and single-marker analyses, regarding the efficiency in selecting and identifying markers or regions associated with QTL. For this, simulated data was used considering six different scenarios, with SNPs being allocated in non-overlapping genomic regions. The second chapter results indicated that for traits with oligogenic inheritance, the WPPA criterion followed by the %var and 𝑃𝑃 𝑖𝑛𝑡 criteria were superior, presenting higher values of detection power, capturing higher percentages of genetic variance and larger areas. The criteria 𝑃𝑃 𝑖𝑛𝑡 and WPPA were considered superior to the others. Also, chapter 3 evaluates the criteria, 𝑃𝑃 𝑖𝑛𝑡 and WPPA, which proved to be superior in chapter 2 together with the single- marker analyses and RHM. However, the detection power and the false positives was assessed considering whether (or not) the inclusion of the dominance effects in the statistical models. For that, simulated data were used in eighteen scenarios with different heritability levels, genetic architecture, and degree of dominance. The results indicated that for the additive effects considering traits with oligogenic genetic architecture, the WPPA, RHM and 𝑃𝑃 𝑖𝑛𝑡 criteria were superior for all analyzed degrees of dominance. For traits with polygenic inheritance, the 𝑃𝑃 𝑖𝑛𝑡 and WPPA criteria can be considered superior. Considering only the effects due to dominance, the criteria WPPA, RHM, single-marker analyses and 𝑃𝑃 𝑖𝑛𝑡 presented relevant results regarding the efficiency measures for the traits controlled by 3 QTL. Keywords: Genomic Regions. Molecular Markers. Bayesian Methods. Genetic Variance.
Descrição
Palavras-chave
Sequenciamento de nucleotídeo, Teoria bayesiana de decisão estatística, Varição genética
Citação
LIMA, Leísa Pires. Métodos aplicados aos estudos de associação genômica via regiões cromossômicas considerando efeitos aditivos e de dominância. 2021. 115 f. Tese (Doutorado em Estatística Aplicada e Biometria) - Universidade Federal de Viçosa, Viçosa. 2021.