Search

For various species, high quality sequences and complete genomes are nowadays available for many individuals. This makes data analysis challenging, as methods need not only to be accurate, but also time efficient given the tremendous amount of data to process. In this article, we introduce an efficient method to infer the evolutionary history of individuals under the multispecies coalescent model in networks (MSNC). Phylogenetic networks are an extension of phylogenetic trees that can contain reticulate nodes, which allow to model complex biological events such as horizontal gene transfer, hybridization and introgression. We present a novel way to compute the likelihood of biallelic markers sampled along genomes whose evolution involved such events. This likelihood computation is at the heart of a Bayesian network inference method called SnappNet, as it extends the Snapp method inferring evolutionary trees under the multispecies coalescent model, to networks. SnappNet is available as a package of the well-known beast 2 software. Recently, the MCMC_BiMarkers method, implemented in PhyloNet, also extended Snapp to networks. Both methods take biallelic markers as input, rely on the same model of evolution and sample networks in a Bayesian framework, though using different methods for computing priors. However, SnappNet relies on algorithms that are exponentially more time-efficient on non-trivial networks. Using simulations, we compare performances of SnappNet and MCMC_BiMarkers. We show that both methods enjoy similar abilities to recover simple networks, but SnappNet is more accurate than MCMC_BiMarkers on more complex network scenarios. Also, on complex networks, SnappNet is found to be extremely faster than MCMC_BiMarkers in terms of time required for the likelihood computation. We finally illustrate SnappNet performances on a rice data set. SnappNet infers a scenario that is consistent with previous results and provides additional understanding of rice evolution.

For various species, high quality sequences and complete genomes are nowadays available for many individuals. This makes data analysis challenging, as methods need not only to be accurate, but also time efficient given the tremendous amount of data to process. In this article, we introduce an efficient method to infer the evolutionary history of individuals under the multispecies coalescent model in networks (MSNC). Phylogenetic networks are an extension of phylogenetic trees that can contain reticulate nodes, which allow to model complex biological events such as horizontal gene transfer, hybridization and introgression. We present a novel way to compute the likelihood of biallelic markers sampled along genomes whose evolution involved such events. This likelihood computation is at the heart of a Bayesian network inference method called SnappNet, as it extends the Snapp method inferring evolutionary trees under the multispecies coalescent model, to networks. SnappNet is available as a package of the well-known beast 2 software. Recently, the MCMC_BiMarkers method, implemented in PhyloNet, also extended Snapp to networks. Both methods take biallelic markers as input, rely on the same model of evolution and sample networks in a Bayesian framework, though using different methods for computing priors. However, SnappNet relies on algorithms that are exponentially more time-efficient on non-trivial networks. Using simulations, we compare performances of SnappNet and MCMC_BiMarkers. We show that both methods enjoy similar abilities to recover simple networks, but SnappNet is more accurate than MCMC_BiMarkers on more complex network scenarios. Also, on complex networks, SnappNet is found to be extremely faster than MCMC_BiMarkers in terms of time required for the likelihood computation. We finally illustrate SnappNet performances on a rice data set. SnappNet infers a scenario that is consistent with previous results and provides additional understanding of rice evolution.

Devant l'afflux d'informations moléculaires, la statistique est un outil indispensable à l'analyse des données issues du génome. Dans ce contexte, l'objectif de ce travail est d'étudier et de proposer des techniques statistiques propres à la détection et à la localisation de loci responsables de la variation d'un caractère quantitatif (Quantitative Trait Loci : QTL). Une première partie est consacrée au selective genotyping (dispositif permettant de réduire les coûts dus au génotypage) et une deuxième partie au génome scan (technique permettant de détecter les QTL en scannant le génome). Dans la première partie, en utilisant des arguments de contiguïté et des calculs d'efficacité on prouve qu'il n'y a aucun gain de puissance à considérer les phénotypes non-extrêmes dans l'analyse statistique, dans le cadre d'un backcross, mais aussi dans un contexte plus général. Si on souhaite génotyper uniquement un pourcentage donné de la population, on montre qu'on doit génotyper le même pourcentage d'individus aux deux extrêmes de la population. On obtient les mêmes conclusions pour un selective genotyping en présence de deux caractères corrélés. Dans le cadre du génome scan, utilisant les lemmes de Le Cam, on présente des résultats asymptotiques sur la distribution du processus de détection sous l'hypothèse nulle d'absence de QTL, sous l'alternative où il existe un seul QTL sur le chromosome et sous l'alternative générale où plusieurs QTL sont présents sur le chromosome. On propose également quelques méthodes adaptées à la carte génétique et permettant d'obtenir le quantile du supremum du processus de détection, indispensable dans la prise de décision. Enfin, on prouve qu'il s'avère inutile d'effectuer des tests sur l'ensemble du chromosome, mais que l'on doit au contraire considérer uniquement quelques positions bien précises., This works aims at studying and proposing statistical techniques dedicated to the detec-tion and the localization of loci responsible for the variation of a quantitative character(Quantitative Trait Loci : QTL). A first part is devoted to selective genotyping (a processallowing to reduce genotyping costs) and the second part to the genome scan (a techniqueallowing to detect QTL when scanning the genome). In the first part, using contiguity ar-guments and efficiency calculations, it is proved that there is no power benefit to considernon extreme phenotypes in the statistical analysis not only in the frame of a backcross,but also in a more general context. When only a given percentage of the populationhas to be genotyped, it is shown that the method is optimal when genotyping the samepercentage of individuals at the two extrema of the population. The same conclusionshold for a selective genotyping with two correlated characters. As far as genome scan isconcerned, using Le Cam’s lemmas, asymptotic results are given on the distribution ofthe detecting process under the null hypothesis of absence of QTL on the chromosome,under the alternative that there is only one QTL and under the general alternative whereseveral QTL are present on the chromosome. These results have been established for po-pulations structured in families of sires. Methods suited to the genetic map are proposedin order to calculate thresholds for the supremum of the process, required for decisionaids. Finally it is shown that testing the whole chromosome is not relevant, only somevery precise positions need to be tested.

Devant l'afflux d'informations moléculaires, la statistique est un outil indispensable à l'analyse des données issues du génome. Dans ce contexte, l'objectif de ce travail est d'étudier et de proposer des techniques statistiques propres à la détection et à la localisation de loci responsables de la variation d'un caractère quantitatif (Quantitative Trait Loci : QTL). Une première partie est consacrée au selective genotyping (dispositif permettant de réduire les coûts dus au génotypage) et une deuxième partie au génome scan (technique permettant de détecter les QTL en scannant le génome). Dans la première partie, en utilisant des arguments de contiguïté et des calculs d'efficacité on prouve qu'il n'y a aucun gain de puissance à considérer les phénotypes non-extrêmes dans l'analyse statistique, dans le cadre d'un backcross, mais aussi dans un contexte plus général. Si on souhaite génotyper uniquement un pourcentage donné de la population, on montre qu'on doit génotyper le même pourcentage d'individus aux deux extrêmes de la population. On obtient les mêmes conclusions pour un selective genotyping en présence de deux caractères corrélés. Dans le cadre du génome scan, utilisant les lemmes de Le Cam, on présente des résultats asymptotiques sur la distribution du processus de détection sous l'hypothèse nulle d'absence de QTL, sous l'alternative où il existe un seul QTL sur le chromosome et sous l'alternative générale où plusieurs QTL sont présents sur le chromosome. On propose également quelques méthodes adaptées à la carte génétique et permettant d'obtenir le quantile du supremum du processus de détection, indispensable dans la prise de décision. Enfin, on prouve qu'il s'avère inutile d'effectuer des tests sur l'ensemble du chromosome, mais que l'on doit au contraire considérer uniquement quelques positions bien précises., This works aims at studying and proposing statistical techniques dedicated to the detec-tion and the localization of loci responsible for the variation of a quantitative character(Quantitative Trait Loci : QTL). A first part is devoted to selective genotyping (a processallowing to reduce genotyping costs) and the second part to the genome scan (a techniqueallowing to detect QTL when scanning the genome). In the first part, using contiguity ar-guments and efficiency calculations, it is proved that there is no power benefit to considernon extreme phenotypes in the statistical analysis not only in the frame of a backcross,but also in a more general context. When only a given percentage of the populationhas to be genotyped, it is shown that the method is optimal when genotyping the samepercentage of individuals at the two extrema of the population. The same conclusionshold for a selective genotyping with two correlated characters. As far as genome scan isconcerned, using Le Cam’s lemmas, asymptotic results are given on the distribution ofthe detecting process under the null hypothesis of absence of QTL on the chromosome,under the alternative that there is only one QTL and under the general alternative whereseveral QTL are present on the chromosome. These results have been established for po-pulations structured in families of sires. Methods suited to the genetic map are proposedin order to calculate thresholds for the supremum of the process, required for decisionaids. Finally it is shown that testing the whole chromosome is not relevant, only somevery precise positions need to be tested.