Resolving deep animal phylogeny with irreversible and unrepeatable genomic changes

解析动物的生命之树对于理解肌肉、肠道和神经系统等复杂特征的起源和进化是至关重要的。基于序列的系统发育学没有达到人们的期望--用新的问题取代了动物系统发生学中的旧问题。不同的数据集和系统发育模型通常会产生强有力的支持但不兼容的树。这其中的一个关键驱动力是通过反向或平行替换，在远相关序列中的给定位置上相同的核苷酸或氨基酸特征状态的收敛进化。这很普遍，因为有几种可能的核苷酸(4)或氨基酸(20)状态可供选择，并且可能产生不正确的树。尽管付出了巨大的努力，但对这一问题和相关问题的适当和易处理的解决方案继续回避着我们，混淆了对动物进化的理解。另一种选择是罕见的基因组变化，如基因、Indels、内含子、基因簇、蛋白质结构等的存在/不存在。这种复杂的特征不太可能是收敛进化的，它们的存在/不存在是一个将物种结合在一起的二元特征。尽管这类方法的潜力和应用越来越多，但已经很明显的是，就像系统基因组学一样，它们也可能被误导；特征丧失可能是收敛的，与原发缺失没有什么区别。为了绕过这一点，我们将建立计算工具，系统地在全基因组范围内识别一类特殊类型的罕见基因组变化，这些变化采用“不可逆”、“不可重复”的特征。这些特征从不缺席，并具有两种同样复杂的保守状态--一种是祖先(存在于非动物和一些动物中)，另一种是派生的(存在于其余动物中)--使得反向或平行进化变得不太可能(例如，基因邻近关系，即“保守但不同”的基因)。这将产生一个新的、自信解决的动物系统学，使我们能够追溯古代前寒武纪海洋中动物特征进化的历史。

Resolving the animal tree of life is essential to understand the origin and evolution of complex traits such as muscles, guts, and a nervous system. Sequence-based phylogenomics has not met expectations- replacing old problems in animal phylogenetics with new ones. Different datasets and phylogenetic models routinely produce strongly supported but incompatible trees. A key driver of this is convergent evolution of the same nucleotide or amino acid character state at a given position in distantly related sequences by reverse or parallel substitutions. This is prevalent, given the few possible alternative nucleotide (4) or amino acid (20) states, and can produce incorrect trees. Despite substantial effort, adequate and tractable solutions to this and related problems continue to evade us, confounding understanding of animal evolution. An alternative is rare genomic changes, e.g. presence/absence of genes, indels, introns, gene clusters, protein domain architecture etc. Such complex characters are very unlikely to have evolved convergently and their presence/absence is a binary character uniting species. Despite the potential and growing use of such methods, it is already clear that like phylogenomics, they can be misled; character loss can be convergent and is indistinguishable from primary absence. To circumvent this, we will build computational tools for systematic genome-wide identification of a special class of rare genomic changes that employ 'irreversible', 'unrepeatable' characters. These characters are never absent and take two equally and highly complex conserved states-one ancestral (present in non-animals and some animals) and one derived (present in the remaining animals)-making reverse or parallel evolution vanishingly unlikely (e.g. gene adjacency relationships, 'conserved-but-different' genes). This will produce a new, confidently-resolved animal phylogeny allowing us to trace the history of animal trait evolution in ancient Precambrian oceans.