Scornavacca Celine, Mayol Joan Carles Pons, Cardona Gabriel
Institut des Sciences de l'Evolution, Université de Montpellier, CNRS, IRD, EPHE 34095 Montpellier Cedex 5, France.
Department of Mathematics and Computer Science, University of the Balearic Islands, E-07122 Palma, Spain.
J Theor Biol. 2017 Apr 7;418:129-137. doi: 10.1016/j.jtbi.2017.01.024. Epub 2017 Jan 20.
In phylogenomics, reconciliations aim at explaining the discrepancies between the evolutionary histories of genes and species. Several reconciliation models are available when the evolution of the species of interest is modelled via phylogenetic trees; the most commonly used are the DL model, accounting for duplications and losses in gene evolution and yielding polynomially-solvable problems, and the DTL model, which also accounts for gene transfers and implies NP-hard problems. However, when dealing with non-tree-like evolutionary events such as hybridisations, phylogenetic networks - and not phylogenetic trees - should be used to model species evolution. Reconciliation models involving phylogenetic networks are still at their early days. In this paper, we propose a new reconciliation model in which the evolution of species is modelled by a special kind of phylogenetic networks - the LGT networks. Our model considers duplications, losses and transfers of genes, but restricts transfers to happen through some specific arcs of the network, called secondary arcs. Moreover, we provide a polynomial algorithm to compute the most parsimonious reconciliation between a gene tree and an LGT network under this model. Our method, when combined with quartet decomposition methods to detect putative "highways" of transfers, permits to refine their analyses by allowing to examine the two possible directions of a highway and even consider combinations of highways.
在系统发育基因组学中,比对旨在解释基因和物种进化历史之间的差异。当通过系统发育树对感兴趣物种的进化进行建模时,有几种比对模型可供使用;最常用的是DL模型,它考虑了基因进化中的复制和丢失,并产生多项式可解问题,以及DTL模型,该模型还考虑了基因转移并意味着NP难问题。然而,当处理诸如杂交等非树状进化事件时,应该使用系统发育网络(而非系统发育树)来对物种进化进行建模。涉及系统发育网络的比对模型仍处于早期阶段。在本文中,我们提出了一种新的比对模型,其中物种进化由一种特殊的系统发育网络——LGT网络进行建模。我们的模型考虑了基因的复制、丢失和转移,但将转移限制为通过网络的一些特定弧(称为二级弧)发生。此外,我们提供了一种多项式算法,用于在此模型下计算基因树和LGT网络之间最简约的比对。我们的方法与四重奏分解方法相结合以检测假定的转移“高速公路”时,通过允许检查高速公路的两个可能方向甚至考虑高速公路的组合,从而能够完善其分析。