The evolution of plant genomes is notable for manifesting a cycle of whole genome doubling, fractionation (gradual loss of redundant genes) and speciation. The thesis is based on a branching process model of the doubling and fractionation process, integrated with a standard model of sequence divergence. The immediate application of this work is to account for the distribution of sequence similarity for duplicate gene pairs, both within plant genomes and between two related plant genomes in terms of a cycle of polyploidization, fractionation and speciation. We derive a mixture distribution for duplicate gene pair similarities generated by speciation and/or repeated episodes of polyploidization. We account not only for the timing of these events in terms of local modes or peaks of the component distributions, but also their volume, or amplitude, and variance. We outline how to infer the parameters of the model. We illustrate with analyses of the distribution of homolog similarities in a number of plant families: Brassicaceae, Solanaceae and Malvaceae. To our knowledge, this is the first method to account for the volume of the component normals of a distribution of similarities, preliminary to an evolutionarily meaningful inference procedure. In addition, we solve the problem of identifying the ploidy level of a series of two or three polyploidizations by invoking the observed and predicted gene triple profiles for each model, i.e., by calculating the probability of the four types of triple with origins in one or the other event, or both.

By increasing genomic size, whole-genome duplication (WGD) is considered a major source of evolutionary innovation and speciation. We examined sequence evolution and expression divergence following WGD in a tetraploid African clawed frog (\textit{Silurana}). We hypothesized that the redundancy generated by WGD might allow for sex-specific and/or tissue-specific divergence, contributing to sexual dimorphism in this frog, and that such changes could be detected at both the expression and sequence levels. We investigated this hypothesis with a transcriptome-based approach, comparing both sexes across brain, heart and liver. We compared molecular evolution and expression divergence of duplicate gene homeologs to singleton genes and to an extant diploid relative, and identified genes with evidence for sex-biased expression. In doing so, we provide evidence for an allopolyploid mechanism of WGD and speciation in \textit{Silurana}. Additionally, we find that female-biased gene expression is more prevalent among duplicate genes than male-biased expression, particularly in brain where expression levels are highest. We similarly identified antagonistically sex-biased homeologs with indication of positive selection. Our results indicate that divergent evolution at both the sequence and expression levels following WGD favors the co-option of female-biased gene expression and may help resolve sexually antagonistic selection in this frog, thereby facilitating the evolution of sexual dimorphism. / Thesis / Master of Science (MSc) / Whole-genome duplication (WGD) is considered a major source of evolutionary innovation and a driver of speciation. By increasing genetic content and introducing redundancy, selective pressures are reduced and paralogous pairs diverge. We investigate how sex and tissue type contribute to duplicate gene divergence following WGD in a tetraploid African clawed frog. We find evidence for sex-dependent variation in sex-biased expression patterns of duplicate genes in brain, heart and liver, and evaluate how molecular evolution of duplicate genes accounts for expression divergence between sexes. This thesis provides a general framework for investigating sex-biased duplicate gene evolution in an amphibious tetrapod.

Genome size in plants varies greatly; in angiosperms this variability can make a 2400- fold difference. This is why the subject of many studies is the attempt to explain these differences in various contexts. One of these is the phylogenetic viewpoint, which studies the relationships between genome size and different taxonomic groups. Many studies have been published so far which explore the plants of the temperate zone; there are however very few studies investigating genome size in tropical plants. For this reason, the subject of this thesis is the analysis of genome size in the pantropical family Costaceae Nakai. This family presently includes seven genera: Monocostus K.Schum., Dimerocostus Kuntze, Costus L., Paracostus C.D.Specht, Chamaecostus C.D.Specht & D.W.Stev, Hellenia Retz. and Tapeinochilos Miq., counting altogether about 300 species. The family is sister to the family Zingiberaceae, which has been notably better explored due to its economic importance. One of the aims of this thesis was a partial revision of the phylogeny published by Specht (2006). Molecular methods were used to analyse the phylogenetic relationships; these methods included classical sequencing of the trnL-F region and next-generation sequencing, where probes designed in advance were used to obtain the targeted genes....

The thesis deals with evolutionary and ecological consequences of edaphic speciation (adaptation to different soil types) and genome duplication (polyploidization), acting in concert. Using a wide range of ecological, karyological and molecular approaches, several hypotheses of general importance have been examined in three model angiosperm systems (ploidy variable species or species aggregates occurring both on and off specific substrates, including serpentines and calcareous soils). In the Knautia arvensis group (Caprifoliaceae) a unique cryptic diploid lineage in central Europe was identified to be restricted to serpentine and limestone outcrops, which served as refugia during environmental changes (forest spread, human impact) in the Holocene. These refugial populations exhibited strong evolutionary potential because they were able to polyploidize and escape beyond the borders of their original edaphically-conditioned refugia owing to hybridization with surrounding widespread homoploid genotypes. Survival of both Knautia cytotypes on serpentine soils was facilitated by their high tolerance to chemical stress factors such as high Ni concentrations and low Ca/Mg ratios. In the Galium pusillum group (Rubiaceae), a striking cytological, ecological, and taxonomic, diversity was revealed in northern and...

Many angiosperms have undergone some series of polyploidization events over the course of their evolutionary history. In these genomes, especially those resulting from multiple autopolyploidization, it may be relatively easy to recognize all the sets of n homeologous chromosomes, but it is much harder, if not impossible, to partition these chromosomes into n subgenomes, each representing one distinct genomic component of chromosomes making up the original polyploid. Thus, if we wish to infer the polyploidization history of the genome, we could make use of all the gene trees inferred from the genes in one set of homeologous chromosomes to construct a consensus tree, but there is no evident way of combining the trees from the different sets because we have no labelling of the chromosomes that is known to be consistent across these sets. We suggest here that lacking a consistent leaf-labelling, the topological structure of the trees may display sufficient resemblance so that a higher level consensus could be revealing of evolutionary history. This would be especially true of the peripheral structures of the tree, likely representing events that occurred more recently and have thus been less obscured by subsequent evolutionary processes. Here, we present a statistical test to assess whether the subgenomes in a polyploid genome could have been added one at a time. The null hypothesis is that the accumulation of chromosomes follows a stochastic process in which transition from one generation to the next is through randomly choosing an edge, and then subdividing this edge in order to link the new internal vertex to a new external vertex. We analyze the probability distributions of a number of peripheral tree substructures, namely leaf- or terminal-pairs, triples and quadruples, arising from this stochastic process, in terms of some exact recurrences. We propose some conjectures regarding the asymptotic behaviours of these distributions. Applying our analysis to a sugarcane genome, we demonstrate that it is unlikely that the accumulation of subgenomes has occurred one at a time in this genome.

Although genome sizes (C-values) are now available for 115 arachnid species (Gregory and Shorthouse [2003] J Hered 94:285-290), the extent of genome amplification (endonuclear DNA replication or polyploidization) accompanying tissue differentiation in this diverse and abundant class of invertebrates remains unknown. To explore this aspect of arachnid development, samples of hemolymph and other tissues were taken from wild-caught specimens as air-dried smears, stained with the Feulgen reaction for DNA, and assayed using both scanning and image analysis densitometry. Cells from midgut diverticula and Malpighian tubules of Argiope and Lycosa (=Pardosa) often showed giant nuclei with 50-100 pg of DNA per nucleus, reflecting at least four cycles of endonuclear DNA replication when compared to the DNA content of hemocytes or sperm from the same specimen. Nuclei with markedly elevated DNA levels also appeared, but far less frequently, in tissue samples from several other arachnid species (Antrodiaetus, Hypochilus, Latrodectus, Liphistus and Loxosceles), but revealed no correlation with differences in somatic cell (2C) genome sizes. Our data show that several DNA classes of polysomatic nuclei regularly arise during tissue differentiation in some species of spiders and may provide an interesting model system for further study of patterns of tissue-specific variation in DNA endoreduplication during development.

The allohexaploid bread wheat (Triticum aestivum L.), evolved through a recent polyploidization event between tetraploid Triticum turgidum L. (AABB) and diploid Aegilops tauschii Coss. (DD), ~8,000 years ago. Contribution of only a subpopulation of Ae. tauschii to hexaploid evolution, followed by domestication and extensive breeding with the objective of higher yield gain and strict end-use quality determining the market classes of wheat, have created a genetic bottleneck. Synthetic hexaploid wheat (SHW) lines are generated to restore the diversity and exploit the genetic resource in the primary gene pool of wheat. However, there are challenges with recovering the phenotypes observed in the parental background in the hexaploid bread wheat in pre-breeding programs. To understand and characterize the barriers in utilizing the progenitor genetic diversity, the transcriptome of four SHW lines and their corresponding tetraploid and diploid parents across ten tissues, totalling to 240 samples, was analysed. The comparison of expression bias of homoeologues present as >18,000 triads (1:1:1) between parental in-silico SHW-like scenarios and SHWs, indicated a large-scale suppression of D subgenome homoeoalleles in SHWs. Tissue-specificity was not observed in the homoeologues of a large proportion of the triads. The next largest fractions were triads where all homoeologues displayed the same tissue-specific expression followed by those where only one of the homoeologues was tissue-specific. Several SHW-tissues showed moderate relationship between tissue-specificity of the homoeologues and expression bias of the corresponding triad. The repression of the genes of the D subgenome was also validated in the differential expression analysis using the entire high-confidence gene set of hexaploid wheat. Qualitative analysis of the transcripts revealed all five splicing events with predominance of retained introns, and more differentially-spliced transcripts were associated with the D homoeoalleles in most SHW-tissue contexts. The introgression patterns of the SHW-C66 into the elite bread wheat cultivar Carberry was analysed using a BC1F5 population. Large introgression of SHW-C66 were found closer to centromeric regions while smaller fragments were present towards the ends of the chromosomes. Correspondingly, the majority of the chromosomes showed higher recombination rates away from the centromere. The donor allele frequency was higher than the expected 25% for BC1F5 population in multiple regions of the A and B subgenomes but not in the D subgenome. In comparison, a preliminary analysis using an elite wheat × elite wheat doubled haploid population showed no subgenome-level variation in recombination rates or donor allele frequencies. In this thesis work, both functional genomic and structural genomic investigations using a set of SHW parents and their derivative population with elite wheat cultivars have unearthed some key patterns that add to the collective knowledge needed to fully exploit genetic resources in broadening the genetic diversity in wheat improvement programs.

La polyploïdisation, e.g. l’augmentation du nombre de jeux de chromosomes chez un organisme, est un phénomène évolutif qui aurait contribué à l’avènement de grandes étapes évolutives, telles que l’apparition de la graine et de la fleur chez les plantes. La polyploïdisation peut être induite par des évènements environnementaux, et notamment par des variations thermiques entraînant la production de gamètes 2n. Il a été montré, chez le rosier, que l’application de températures élevées (dès 30°) lors de la gamétogénèse mâle, induisait une forte production de diplogamètes ainsi qu’une forte répression de l’expression du gène RhPS1, orthologue du gène AtPS1 d’A. thaliana. AtPS1 est impliqué dans la formation de diplogamètes, dont le mutant perte de fonction présente une forte production de diplogamètes mâles, issus d’anomalies méiotiques en Métaphase II. L’objectif de cette thèse est d’apporter des éléments permettant de mieux connaitre et de caractériser la fonction de RhPS1, par (i) l’analyse des phénotypes de lignées de rosier présentant une diminution de ces transcrits, (ii) en recherchant ces interactants protéiques potentiels et (iii) en recherchant les gènes potentiellement régulés par RhPS1. Les résultats de cette thèse montrent que RhPS1 aurait une fonction similaire à AtPS1 chez A. thaliana, qui induit, lors d’une baisse d’expression, des anomalies méiotiques uniquement lors des étapes de métaphase II ou lors de la transition Anaphase I/Métaphase II. Même si aucun interactant protéique n’a pu être trouvé, quelques gènes potentiellement régulés par RhPS1 ont pu être identifiés. / Polyploidization, e.g. the increase in the number of chromosomes sets in an organism, is a phenomenon that would have contributed to the advent of major evolutionary steps such as the emergence of the seed and the flower in plants.Polyploidization can be induced by environmental clues, more particularly by thermal variations resulting in the production of 2n. It has been shown, in roses, that the application of high temperatures (from 30°C) during male gametogenesis leads to a high production of diplogametes as well as a high repression of the expression of the gene RhPS1, orthologue of the A. Thaliana AtPS1 gene. AtPS1 is involved in the formation of diplogametes, whose loss of function mutant has a strong production of male diplogametes, resulting from meiotic abnormalities. This work aims at characterizing or providing elements allowing for better comprehension of the RhPS1 function, by (i) the analysis of the phenotypes of rose lines showing a decrease of these transcripts, (ii) by looking for these potential protein interactors and (iii) investigating genes potentially regulated by RhPS1. The results of this work show that RhPS1 has a similar function to AtPS1 in A. thaliana. Thus, when this gene has a lower expression, meiotic abnormalities are only observed during the metaphase II stages or during the transition of Anaphase I / Metaphase II. Even though no protein interactant could be found, some genes potentially regulated by RhPS1 have been identified.

Polyploidization is exceptionally cosmopolitan in plants. One common observation is that polyploids inhabit larger geographic distributions than their diploid progenitors. Differences in distribution between cytotypes are largely attributed to differences in mating system and ecological breadth among cytotypes. In Crataegus series Douglasianae, allopolyploids have larger ranges than their diploid progenitors. Range size increase is coupled with a shift to predominant asexuality in polyploids. This thesis explores 2 additional hypotheses that may contribute to differences in distribution among cytotypes: ecological breadth and dispersal ability. We find evidence that tetraploid C. douglasii occurs in a wider range of habits and has a greater dispersal ability than diploid C. suksdorfii. Overall, we suggest that differences in mating system, ecological breadth, and dispersal ability have contributed individually and collaboratively to differences in distribution among cytotypes of Crataegus series Douglasianae. Large ranges in polyploids may help maintain cytotype diversity by providing buffering capacity against demographic stochasticity.

Polyploidization is exceptionally cosmopolitan in plants. One common observation is that polyploids inhabit larger geographic distributions than their diploid progenitors. Differences in distribution between cytotypes are largely attributed to differences in mating system and ecological breadth among cytotypes. In Crataegus series Douglasianae, allopolyploids have larger ranges than their diploid progenitors. Range size increase is coupled with a shift to predominant asexuality in polyploids. This thesis explores 2 additional hypotheses that may contribute to differences in distribution among cytotypes: ecological breadth and dispersal ability. We find evidence that tetraploid C. douglasii occurs in a wider range of habits and has a greater dispersal ability than diploid C. suksdorfii. Overall, we suggest that differences in mating system, ecological breadth, and dispersal ability have contributed individually and collaboratively to differences in distribution among cytotypes of Crataegus series Douglasianae. Large ranges in polyploids may help maintain cytotype diversity by providing buffering capacity against demographic stochasticity.