Hilu Khidir W, Borsch Thomas, Müller Kai, Soltis Douglas E, Soltis Pamela S, Savolainen Vincent, Chase Mark W, Powell Martyn P, Alice Lawrence A, Evans Rodger, Sauquet Hervé, Neinhuis Christoph, Slotta Tracey A B, Rohwer Jens G, Campbell Christopher S, Chatrou Lars W
Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 USA;
Am J Bot. 2003 Dec;90(12):1758-76. doi: 10.3732/ajb.90.12.1758.
Plastid matK gene sequences for 374 genera representing all angiosperm orders and 12 genera of gymnosperms were analyzed using parsimony (MP) and Bayesian inference (BI) approaches. Traditionally, slowly evolving genomic regions have been preferred for deep-level phylogenetic inference in angiosperms. The matK gene evolves approximately three times faster than the widely used plastid genes rbcL and atpB. The MP and BI trees are highly congruent. The robustness of the strict consensus tree supercedes all individual gene analyses and is comparable only to multigene-based phylogenies. Of the 385 nodes resolved, 79% are supported by high jackknife values, averaging 88%. Amborella is sister to the remaining angiosperms, followed by a grade of Nymphaeaceae and Austrobaileyales. Bayesian inference resolves Amborella + Nymphaeaceae as sister to the rest, but with weak (0.42) posterior probability. The MP analysis shows a trichotomy sister to the Austrobaileyales representing eumagnoliids, monocots + Chloranthales, and Ceratophyllum + eudicots. The matK gene produces the highest internal support yet for basal eudicots and, within core eudicots, resolves a crown group comprising Berberidopsidaceae/Aextoxicaceae, Santalales, and Caryophyllales + asterids. Moreover, matK sequences provide good resolution within many angiosperm orders. Combined analyses of matK and other rapidly evolving DNA regions with available multigene data sets have strong potential to enhance resolution and internal support in deep level angiosperm phylogenetics and provide additional insights into angiosperm evolution.
利用简约法（MP）和贝叶斯推断法（BI），对代表所有被子植物目374个属以及12个裸子植物属的质体matK基因序列进行了分析。传统上，进化缓慢的基因组区域在被子植物深层系统发育推断中更受青睐。matK基因的进化速度大约是广泛使用的质体基因rbcL和atpB的三倍。MP树和BI树高度一致。严格合意树的稳健性超过了所有单个基因分析，仅与基于多基因的系统发育相当。在解析出的385个节点中，79%得到了较高的自展值支持，平均为88%。无油樟是其余被子植物的姐妹群，其次是睡莲科和木兰藤目的一个分支。贝叶斯推断将无油樟+睡莲科解析为其余类群的姐妹群，但后验概率较弱（0.42）。MP分析显示，与木兰藤目构成三分关系的是真木兰类、单子叶植物+金粟兰目，以及金鱼藻+真双子叶植物。matK基因对基部真双子叶植物提供了迄今为止最高的内部支持，并且在核心真双子叶植物中，解析出一个包括小檗叶科/毒鼠子科、檀香目，以及石竹目+菊类植物的冠群。此外，matK序列在许多被子植物目中提供了良好的分辨率。将matK与其他快速进化的DNA区域与现有的多基因数据集进行联合分析，在增强被子植物深层系统发育分辨率和内部支持方面具有强大潜力，并能为被子植物进化提供更多见解。
