What drives evolutionary diversification in a biodiversity hotspot? A `next generation´ phylogenetic approach comparing three clades of the Cape Floristic Region´s mega-genus Erica

南非开普区（Cape Floristic Region， CFR）是一个植物研究热点，是研究生物多样性的理想系统。CFR的总面积很小，但在那里发现的植物物种数量却高得不成比例，其中惊人的70%是世界上其他地方所没有的。最大的海角分支Erica属（Erica科）仅占植物群的7%，使其成为了解进化多样化原因的关键分类群。在我目前的dfg资助项目的更新中，我建议测试一些最可能的因素，这些因素被认为是CFR植物多样性高的潜在原因，主要是与地理范围和传粉媒介的变化有关。从最初的目标延伸，我们将比较三组密切相关的Erica物种：brbricata /coccinea-枝（数据收集于当前项目），abietina/viscaria-枝和E. plukenetii。这些分支共包括c. 34种，通常具有高度地方性的地理分布。它们的特点是花的宏观形态特征的变化，这些特征通常与不同媒介的传粉有关。这些分支中最近出现的物种和种群的分化意味着进化事件的系统发育证据，比如分散，可能会被随后的事件所掩盖。然而，当比较密切相关的个体和分支时，网状过程，如杂交和聚结的随机性会导致基因树之间的相当大的差异，这在我们的brbricata /coccinea-clade数据集中是显而易见的。只有在这种情况下，如果我们对大量有意义的独立基因树进行采样，并适当地模拟其差异背后的过程，才有可能进行准确的系统发育推断。为此，我建议使用我们的erica优化杂交捕获和高通量测序管道，为这两个额外进化支的72个代表性标本中的每一个获得132个，高达c. 7 Kb长的DNA序列标记。为了改进种群采样，我们将使用多重PCR方法获得多达192个进一步样本的标记子集。通过用基于聚结的方法分析数据，我们可以推断出有意义的、潜在的网状物种树。然后，所有三个分支的物种树可以与进一步基于系统发育的推断方法一起使用，作为对地理范围和特征进化相互竞争的假设的独立测试。用这些数据验证的合理假设包括异域物种形成模型和传粉媒介驱动物种形成的Grant-Stebbins模型。通过这样做，我们可以深入了解在开普角和其他地方驱动物种分化的过程。

The Cape Floristic Region (CFR) of South Africa is a botanical hotspot and represents an ideal system with which to investigate biological diversity. The total area of the CFR is small but the number of plant species found there is disproportionally high, and an astonishing 70% of them are found nowhere else in the world. The largest Cape clade, the genus Erica (Ericaceae), alone accounts for 7% of the flora, making it a key taxon for understanding the causes of evolutionary diversification. In a renewal of my current DFG-funded project, I propose to test a number of the most likely factors proposed as potential causes of the high plant diversity of the CFR, principally those relating to shifts in geographic range and pollinators. Extending from the original aims, we would compare three groups of closely related Erica species: the imbricata/coccinea-clade (data collected in current project), the abietina/viscaria-clade, and E. plukenetii. The clades comprise a total of c. 34 species, with often highly localised geographic distributions. They are each characterised by variation in floral macro-morphological characters that are generally related to pollination by different vectors. The recent divergence of the species and populations in these clades means that phylogenetic evidence for evolutionary events, such as dispersals, may remain unmasked by subsequent events. However, reticulate processes, such as hybridisation, and coalescent stochasticity can result in considerable differences between gene trees when comparing closely related individuals and clades, as is apparent in our imbricata/coccinea-clade dataset. Accurate phylogenetic inference is only possible under these circumstances if we sample numerous, meaningfully resolved independent gene trees and model the processes underlying their differences appropriately. To this end, I propose to use our Erica-optimised hybridisation capture and high-throughput sequencing pipeline to obtain 132, up to c. 7 Kb long DNA sequence markers for each of 72 representative specimens of the two additional clades. For improved sampling of populations, we will use a multiplex PCR approach to obtain a subset of the markers for up to 192 further samples. By analysing the data with coalescence-based methods we can infer meaningful, potentially reticulate, species trees. The species trees of all three clades can then be used with further phylogeny-based inference methods as independent tests of competing hypotheses of geographic range and character evolution. Plausible hypotheses to be tested with this data include allopatric speciation models and the Grant-Stebbins model for pollinator driven speciation. In so doing we can yield profound insight into the processes driving the divergence of species, both in the Cape, and in general.