Admixture, the genetic exchange between populations, is a topic of great interest in evolutionary biology. It has been found using observational approaches, and appears to be a common phenomenon in the history of many species across the tree of life. The advent of genomics has provided new ways to detect admixture, quantify the amount of introgressed DNA and pinpoint it in genomes. As a consequence, the field is in the midst of an age of discovery. This is the case for mammals,^[1] as well as more established systems for speciation and gene flow such as birds and fishes. It is becoming clear that admixture is common rather than an exception.

In this issue of BioEssays, Ottenburghs^[2] puts forward the idea that so‐called “ghost” admixture events are emerging as an important topic. These are admixture events in which the source population is unknown, but genetic data allow us to study them. So far, most attention has been on comparisons between members of current groups whose ancestors have exchanged genes in the past (such as brown bears and polar bears), or with extinct groups for which ancient DNA is available (such as Neanderthals and modern humans). However, in cases where neither descendants of the source population exist, nor data from ancient DNA can be obtained, traces of gene flow are still buried in the genomes of present‐day individuals. Ottenburghs points out that new data and methods to discover these will make it possible to add branches to phylogenies purely from a genomic perspective, and even make inferences about the biology of the respective organisms.

In some cases, admixture from unknown sources has been found by deep divergence in plastid DNA. However, these contain very limited information for biological inferences, meaning that there is much greater potential in methods that screen multiple whole genomes for highly divergent haplotypes: this approach allows access to partial genomes of extinct populations. The resurrection of such ancient lineages from present‐day individuals is a dynamic research area, where new methods are emerging fast, each of them with inherent limitations. Often, these methods are being developed around cases where the “ghosts” have been sequenced, especially archaic hominin forms.

Ultimately, demographic modeling is necessary to distinguish archaic admixture from other processes, an example being undiscovered population structure. A limitation is that each model is an approximation of a likely complex biological reality, where population separation and admixture pulses are only simplifications of long‐term processes. Hence, it is obviously desirable to make use of as large a fraction of the genome, and as many individuals, as possible. Because timescales of evolutionary change can vary by orders of magnitude across clades, there are vastly differing expectations for parameters of demographic models. Thus, the evaluation of these models remains a major challenge in the field.

Beyond describing the occurrence of admixture events, excavating the genomes of “ghosts” has great potential for guiding the understanding of speciation and adaptation of the surviving lineages. There is plenty of evidence that gene flow introduced variation that facilitated adaptation, good examples being the immune system and other traits in humans and other species. On the other hand, introgression deserts, where hybrid incompatibilities have left their footprint in the genome, may shed light on what drives speciation. Finally, the field of inferring phenotypes from genotypes is rapidly developing, providing tools and databases that can be used to study the influence on phenotypic traits. Ottenburghs emphasizes that traits such as wing patterns in butterflies can deviate from the overall phylogenetic tree on account of past transfer between species. Finally, the detection of ghost admixture allows inferences on traits in these extinct ghost populations in the absence of fossils or other types of specimen.

This is a dynamically evolving field, with broad implications for how we understand speciation: indeed, it may even imply “the abolition of species” in a strict conceptual sense. In line with Ottenburghs article, I expect exciting and surprising insights on this topic in the next few years.

种群之间的遗传交换，即混合物，是进化生物学中非常感兴趣的话题。它是通过观察方法发现的，似乎是整个生命树上许多物种历史上的普遍现象。基因组学的出现提供了检测混合物的新方法，量化了渗入的DNA量并在基因组中进行了精确定位。结果，该领域处于发现时代之中。哺乳动物就是这种情况，^[1]以及鸟类和鱼类等更加成熟的物种形成和基因流动系统。显然，混合是常见的，而不是例外。

在这个问题上BIOESSAYS，Ottenburghs ^[2]提出了一个想法，即所谓的“鬼”混合事件正在成为一个重要的话题。这些是混合事件，其中来源种群未知，但是遗传数据使我们能够研究它们。到目前为止，最受关注的是那些祖先在过去交换过基因的当前群体（例如棕熊和北极熊）之间的比较，或者与那些拥有古代DNA的灭绝群体（例如尼安德特人和现代人类）之间的比较。 。但是，如果既没有源种群的后代，也无法获得来自古代DNA的数据，则基因流的痕迹仍然埋在当今个体的基因组中。奥滕伯格（Ottenburghs）指出，发现这些数据的新数据和方法将有可能纯粹从基因组的角度为系统发育增加分支，

在某些情况下，通过质体DNA的深层差异发现了未知来源的混合物。但是，这些方法所包含的生物学推论信息非常有限，这意味着在筛选多个完整基因组以寻找高度单倍型的方法中，潜力更大：这种方法可以访问已灭绝种群的部分基因组。从当今个体中复活这种古老的血统是一个动态的研究领域，新方法正在迅速出现，每种方法都有其固有的局限性。通常，这些方法是围绕“鬼魂”已被测序的情况开发的，尤其是古人味素形式。

最终，人口模型是区分旧式掺混物与其他过程的必要条件，例如未发现的种群结构。局限性在于每个模型都是可能的复杂生物现实的近似，其中种群分离和混合脉冲只是长期过程的简化。因此，显然希望利用基因组的很大一部分，并尽可能多地使用个体。由于进化变化的时间尺度在进化枝之间可能会变化几个数量级，因此对人口统计模型参数的期望存在很大差异。因此，对这些模型的评估仍然是该领域的主要挑战。

除了描述混合事件的发生以外，挖掘“鬼魂”的基因组还具有巨大的潜力，可以指导对物种的理解和对尚存谱系的适应。有大量证据表明基因流引入了有助于适应的变异，这是人类和其他物种的免疫系统和其他特征的很好例证。另一方面，杂种不相容性已在基因组中留下痕迹的渗入沙漠可能会揭示驱动物种形成的原因。最后，从基因型推断表型的领域正在迅速发展，提供了可用于研究对表型性状影响的工具和数据库。奥滕伯格强调指出，由于过去物种之间的转移，蝴蝶的翅膀图案等性状可能会偏离整个系统发育树。最后，在没有化石或其他类型标本的情况下，鬼混物的检测可以推断这些已灭绝的鬼群的性状。

这是一个动态发展的领域，对我们对物种的理解有着广泛的含义：实际上，它甚至可能意味着严格意义上的“物种灭绝”。与Ottenburgh的文章一致，我希望在接下来的几年中对该主题有令人兴奋和令人惊讶的见解。