A phylogenetic hypothesis of relationships among 33 species of stalk-eyed flies was generated from a molecular data set comprising three mitochondrial and three nuclear gene regions. A combined analysis of all the data equally weighted produced a single most-parsimonious cladogram with relatively strong support at the majority of nodes. The phylogenetic utility of different classes of molecular data was also examined. In particular, using a number of different measures of utility in both a combined and separate analysis framework, we focused on the distinction between mitochondrial and nuclear genes and between faster-evolving characters and slower-evolving characters. For the first comparison, by nearly any measure of utility, the nuclear genes are substantially more informative for resolving diopsid relationships than are the mitochondrial genes. The nuclear genes exhibit less homoplasy, are less incongruent with one another and with the combined data, and contribute more support to the combined analysis topology than do the mitochondrial genes. Results from the second comparison, however, provide little evidence of a clear difference in utility. Despite indications of rapid divergence and saturation, faster-evolving characters in both the nuclear and mitochondrial data sets still provide substantial phylogenetic signal. In general, inclusion of the more rapidly evolving data consistently improves the congruence among partitions.

中文翻译：

---

不同类型的分子数据的系统发生学效用，可用于推断茎状蝇（拟蝇科）之间的进化关系。

从包括三个线粒体和三个核基因区域的分子数据集，产生了33种带茎苍蝇之间关系的系统发育假说。对所有数据均等加权的综合分析产生了一个最简约的单幅图，在大多数节点上都具有相对强大的支持。还检查了不同类别的分子数据的系统发生学效用。特别是，在组合分析和单独分析框架中使用多种不同的效用度量方法，我们集中于线粒体和核基因之间以及快速发展的特征和缓慢发展的特征之间的区别。对于第一个比较，通过几乎任何实用性度量，与线粒体基因相比，核基因对于解决双视关系的信息要多得多。与线粒体基因相比，核基因展示出更少的同质性，彼此之间以及与组合数据之间的不一致程度更低，并且为组合分析拓扑提供了更多支持。然而，第二次比较的结果几乎没有提供效用明显差异的证据。尽管有迅速发散和饱和的迹象，但核和线粒体数据集中发展较快的特征仍然提供了重要的系统发生信号。通常，包含更快发展的数据始终可以提高分区之间的一致性。几乎没有提供效用明显差异的证据。尽管有迅速发散和饱和的迹象，但核和线粒体数据集中发展较快的特征仍然提供了重要的系统发生信号。通常，包含更快发展的数据始终可以提高分区之间的一致性。几乎没有提供效用明显差异的证据。尽管有迅速发散和饱和的迹象，但核和线粒体数据集中发展较快的特征仍然提供了重要的系统发生信号。通常，包含更快发展的数据始终可以提高分区之间的一致性。