Search

MotivationThe reconstruction of ancestral genetic sequences from the analysis of contemporaneous data is a powerful tool to improve our understanding of molecular evolution. Various statistical criteria defined in a phylogenetic framework can be used to infer nucleotide, amino-acid or codon states at internal nodes of the tree, for every position along the sequence. These criteria generally select the state that maximizes (or minimizes) a given criterion. Although it is perfectly sensible from a statistical perspective, that strategy fails to convey useful information about the level of uncertainty associated to the inference.ResultsThe present study introduces a new criterion for ancestral sequence reconstruction, the minimum posterior expected error (MPEE), that selects a single state whenever the signal conveyed by the data is strong, and a combination of multiple states otherwise. We also assess the performance of a criterion based on the Brier scoring scheme which, like MPEE, does not rely on any tuning parameters. The precision and accuracy of several other criteria that involve arbitrarily set tuning parameters are also evaluated. Large scale simulations demonstrate the benefits of using the MPEE and Brier-based criteria with a substantial increase in the accuracy of the inference of past sequences compared to the standard approach and realistic compromises on the precision of the solutions returned.Availability and implementationThe software package PhyML (https://github.com/stephaneguindon/phyml) provides an implementation of the Maximum A Posteriori (MAP) and MPEE criteria for reconstructing ancestral nucleotide and amino-acid sequences.

Different species have different genome organization. Whether it be the karyotype or gene order, these differences are seen even with relatively close species like Human and Mouse. This is caused by the chromosomal rearrangement. Infererence of rearrangement scenarios that transform one present-day species into another can give insight into evolutionary states, the ancestral genome being one of the intermediates of the true scenario.The chromosomal rearrangements are violent biological events for the cell. Indeed, numerous mechanisms are present to stop the cell cycle when the genome sequence is altered. Moreover, rearrangements can be the source of aberrant phenotypes, which are probably unfavorable for the carrier. With all that, it seams reasonable to assume the rearrangement scenarios are parsimonious.However, it is accepted that this criterion alone is not sufficient to efficiently build the evolutionary history of the genomes. Indeed, for whatever model we choose, the number of scenario is exponential in the number of rearrangements. Another biological constraint is needed. The spatial structure of the chromatin could be an essential missing criterion. It has been shown in vitro that when a double-stranded break of the DNA is non-homologously repaired, the strand used for repairing is close in space to the breakpoint. Our hypothesis is that the closer the breakpoints are in space, the more probable they are to participate in a rearrangement. This hold on genomics analysis of somatic cells, and between species. Let's name that hypothesis the locality hypothesis.We proposed a method to use the structural information in order to prioritize the rearrangements scenarios. The Hi-C data were the structural information that allowed us to apply our method to scenarios between D. melanogaster and D. yakuba.This results led us to ask whether the chromatin structure could evolve by itself. Then, it could be used as a phylogenetic mark. This idea is related to previous results showing the conservation of topological domains between species.This question seams to be new, and could open a new line of investigation. If the chromatin structure holds a phylogenetical signal, it becomes possible to ask ourselves about the mechanisms that occur during the selection, or if it is possible for the ancestral state to be inferred. Then, it could even be possible to compare the evolution of the sequence with the one of the chromatin structure.Thus, we defined a distance between genome structures, based on the comparison of contacts between orthologous loci. We applied this distance to a set of six species, including the Human, the Mouse and four Drosophila. This result confirms the presence of a phylogenetic signal in the spatial structure of the genomes. They also showed that we're in need for efficient methods to compare contacts data between species.; Entre espèces, les génomes présentent des différences dans leur organisation, que ce soit au niveau du caryotype ou de l'ordre des gènes. Ceci reste vrai même entre espèces relativement proches comme l'humain et la souris, et est du aux réarrangements chromosomiques. Reconstruire l'histoire évolutive d'une lignée revient donc à déterminer des scénarios de réarrangements qui transforment un génome actuel en un autre. Le génome ancestral se trouve alors être l'un des états intermédiaires atteint par l'un de ces scénarios.Les réarrangements chromosomiques sont des évènements biologiques violents pour la cellule. En effet, de nombreux mécanismes moléculaires ont pour fonction de stopper le cycle cellulaire dans le cas où le génome aurait été altéré. De plus, les réarrangements peuvent être à l'origine de phénotypes aberrants, et donc probablement désavantageux pour leur porteur. Au vu de tout cela, il paraît raisonnable de poser l'hypothèse selon laquelle les scénarios de réarrangements sont parcimonieux.Cependant, il est admis que ce seul critère ne permet pas de reconstruire efficacement l'histoire évolutive des génomes. En effet, quelque soit le modèle utilisé pour générer les scénarios, leur nombre est exponentiel en le nombre de réarrangements. Une autre contrainte biologique doit donc être ajoutée. La conservation de la structure spatiale de la chromatine pourrait être un critère manquant essentiel. Il a été montré in vitro que lors d'une cassure double-brin suivie d'une réparation non-homologue, le brin utilisé pour la réparation se situe spatialement proche de la cassure. Notre hypothèse est donc que les points de cassures qui sont proches en 3D ont plus probablement participé à des réarrangements que les autres. Cela est appuyé par des analyses génomiques sur des cellules somatiques et entre espèces. Nommons cette hypothèse: l'hypothèse de localité.Notre approche a été de proposer une méthode pour utiliser l'information structurale afin de prioriser les scénarios de réarrangements. Les données de Hi-C ont été l'information structurale qui nous a permis d'appliquer la méthode aux scénarios entre D. melanogaster et D. yakuba.Ces résultats nous ont ensuite menés à nous demander si la structure de la chromatine ne pouvait pas elle-même évoluer. Elle serait alors susceptible d'être considérée comme un caractère phylogénétique. Cette idée est appuyée par d'autres résultats montrant la conservation de domaines topologiques entre espèces.Cette question ne semble pas avoir été posée auparavant. Elle est pourtant très intéressante car elle permet d'ouvrir tout un champ d'étude. En effet, si la structure de la chromatine porte un signal phylogénétique, alors il devient possible de s'interroger sur les mécanismes en œuvre lors de la sélection, ou sur la possibilité de reconstruire l'état ancestral de cette structure. Par la suite, il serait même possible de comparer l'évolution de la séquence et celle de la structure de la chromatine.Nous avons ainsi défini une distance entre les structures des génomes, basée sur la comparaison des contacts entre loci orthologues. Nous l'avons appliquée à une ensemble de six espèces comprenant l'humain, la souris et quatre drosophiles. Ces résultats confirment la présence d'un signal phylogénétique dans la structure spatiale des génomes. Ils mettent également en lumière l'intérêt de la mise en place de méthodes permettant de comparer efficacement des données de contacts entre espèces.

The reconstruction of ancestral genetic sequences from the analysis of contemporaneous data is a powerful tool to improve our understanding of molecular evolution. Various statistical criteria defined in a phylogenetic framework can be used to infer nucleotide, aminoa-cid or codon states at internal nodes of the tree, for every position along the sequence. These criteria generally select the state that maximises (or minimises) a given criterion. Although it is perfectly sensible from a statistical perspective, that strategy fails to convey useful information about the level of uncertainty associated to the inference. The present study introduces a new criterion for ancestral nucleotide reconstruction that selects a single state whenever the signal conveyed by the data is strong, and a combination of multiple states otherwise. Simulations demonstrate the benefit of this approach with a substantial increase in the accuracy of ancestral sequence reconstruction without significantly compromising on the precision of the solutions returned.

International audience; Genome architecture can be drastically modified through a succession of large-scale rearrangements. In the quest to infer these rearrangement scenarios, it is often the case that the parsimony principal alone does not impose enough constraints. In this paper we make an initial effort towards computing scenarios that respect chromosome con-formation, by using Hi-C data to guide our computations. We confirm the validity of a model – along with optimization problems Minimum Local Scenario and Minimum Local Parsimonious Scenario – developed in previous work that is based on a partition into equivalence classes of the adjacencies between syntenic blocks. To accomplish this we show that the quality of a clustering of the adjacencies based on Hi-C data is directly correlated to the quality of a rearrangement scenario that we compute between Drosophila melanogaster and Drosophila yakuba. We evaluate a simple greedy strategy to choose the next rearrangement based on Hi-C, and motivate the study of the solution space of Minimum Local Parsimonious Scenario.

Genome architecture can be drastically modified through a succession of large-scale rearrangements. In the quest to infer accurate ancestral rearrangement scenarios, it is often the case that parsimony principal alone does not impose enough constraints. Thus, the current challenge is to consider more biological information in the inference process. In previous work, we introduced a model for such a task, based on a partition into equivalence classes of the adjacencies between genes. Such a partition is amenable to the representation of spacial constraints as given by Hi-C data. A major open question is the validity of such a model. In this note, we show that the quality of a clustering of the adjacencies based on Hi-C data is directly correlated to the quality of a rearrangement scenario that we compute between Drosophila melanogaster and D. yakuba.