Duplication, Rearrangement and Reconciliation: A Follow-Up 13 Years Later

Abstract

Abstract. The evolution of genomes can be studied at at least three different scales: the nucleotide level, accounting for substitutions and indels, the gene level, accounting for gains and losses, and the genome level, accounting for rearrangements of chromosome organization. While the nucleotide and gene levels are now often integrated in a single model using reconciled gene trees, very little work integrates the genome level as well, and considers gene trees and gene orders simultaneously. In a seminal book chapter published in 2000 and entitled "Duplication, Rearrangement and Reconciliation", Sankoff and El-Mabrouk outlined a general approach, making a step in that direction. This avenue has been poorly exploited by the community for over ten years, but recent developments allow the design of integrated methods where phylogeny informs the study of synteny and vice-versa. We review these developments and show how this influence of synteny on gene tree construction can be implemented.

ASTRAL-Pro: quartet-based species tree inference despite paralogy

“…There is a large literature on various ways of tagging and rooting gene trees [e.g., [64][65][66], including other penalties for the duplication and loss events (e.g., there is a suggestion of losses having half the penalty of duplications [67]). It may also be possible to improve tagging of gene trees using probabilistic orthology inference [68,69] or using synteny information [70,71]. However, these methods often require a species tree.…”

Synteny-guided resolution of gene trees clarifies the functional impact of whole genome duplications

“…Genome evolution operates through three major mechanisms: nucleotide substitutions, gene duplications and losses, and genomic rearrangements. To date, integrating these different evolutionary events into a unified framework remains an open challenge (Chauve et al 2013). 455…”

“…Synteny conservation has the potential to neatly complement sequence similarity in gene evolution studies, because gene order evolves via independent mechanisms, and is more resilient to saturation at deep evolutionary times (Rokas and Holland 2000). Genome 460 organisation information still remains difficult to incorporate in gene phylogenies, largely due to the lack of well-supported evolutionary models (Chauve et al 2013) and the need for contiguous genome assemblies. The most notable effort to use extant synteny to correct gene trees in a general context showed mixed results (ParalogyCorrector within the RefineTree framework (Lafond et al 2013;Noutahi et al 2016)).…”

Assessing the Robustness of Parsimonious Predictions for Gene Neighborhoods from Reconciled Phylogenies: Supplementary Material

“…This is consistent with the fact that for rare evolutionary events such as genome rearrangements, a parsimony approach is relevant, especially when it can be complemented by efficient algorithms to explore slightly sub-optimal solutions, such as DeClone, and to explore the parameter space. In terms of direct applications of the method developed here and in [6], gene-tree based reconstruction of ancestral gene orders comes to mind [5]; more precisely, ancestral adjacencies could be determined and scored using a mixture of their Boltzmann probability (that can be computed efficiently using DeClone) and robustness to changes of the cost scheme, and conflicts could be cleared out independently and efficiently for each ancestral species using the algorithm of [12] for example. An interesting observation is that even the set of ancestral adjacencies that are universally-parsimonious and robust to changes in the cost scheme contains a significant number of adjacencies participating in syntenic conflict.…”

“…They defined a model for the evolution of gene adjacencies within a species phylogeny, together with an efficient dynamic programming (DP) algorithm, called DeCo, to compute parsimonious evolutionary histories that minimize the total cost of gene adjacencies gain and break, for a given cost scheme associating a cost to each of these two events. Reconstructing evolutionary scenarios for syntenic characters is an important step towards more comprehensive models of genome evolution, going beyond classical sequence/ content frameworks, as it implicitly integrates genome rearrangements [5]. Application of such methods include the study of genome rearrangement rates and the reconstruction of ancestral gene order.…”