文獻出處: Zwick A., C. J. Regier, C. Mitter and M. P. Cummings. 2010. Increased gene sampling yields robust support for higher-level clades within Bombycoidea (Lepidoptera). Systematic Entomology. DOI: 10.1111/j.1365-3113.2010.00543.x
2008年由美國馬里蘭大學生物科技研究所Jerome C. Regier與現德國斯圖加特自然史博物館的Andreas Zwick博士等發表了針對蠶蛾總科與枯葉蛾總科親緣關係重建的研究。Regier等人選用50個蠶蛾總科、近緣枯葉蛾科與其他大鱗翅分類群，並取5段蛋白質轉錄基因序列(約6600鹼基對)進行親緣關係重建，然而此研究中蠶蛾總科並未獲得高支持度的單系性。
資料組 (A)- 50個分類群(非完全等同Regier等人2008年選用分類群)/5段序列
資料組 (B)- 50個分類群(同上述)/25段序列
資料組 (C)- 24個分類群/5段序列
(2) 擴增部份分類群之基因序列數量，造成其他分類群相對產生了missing data，然而卻並不會造成顯著的非親緣關係偏誤(nonphylogenetic bias)
This study has as its primary aim the robust resolution of higher-level relationships within the lepidopteran superfamily Bombycoidea. Our study builds on an earlier analysis of five genes (∼6.6 kbp) sequenced for 50 taxa from Bombycoidea and its sister group Lasiocampidae, plus representatives of other macrolepidoteran superfamilies. The earlier study failed to yield strong support for the monophyly of and basal splits within Bombycoidea, among others. Therefore, in an effort to increase support specifically for higher-level nodes, we generated 11.7 kbp of additional data from 20 genes for 24 of 50 bombycoid and lasiocampid taxa. The data from the genes are all derived from protein-coding nuclear genes previously used to resolve other lepidopteran relationships. With these additional data, all but a few higher-level nodes are strongly supported. Given our decision to minimize project costs by augmenting genes for only 24 of the 50 taxa, we explored whether the resulting pattern of missing data in the combined-gene matrix introduced a nonphylogenetic bias, a possibility reported by others. This was achieved by comparing node support values (i.e. nonparametric bootstrap values) based on likelihood and parsimony analyses of three datasets that differ in their number of taxa and level of missing data: 50 taxa/5 genes (dataset A), 50 taxa/25 genes (dataset B) and 24 taxa/25 genes (dataset C). Whereas datasets B and C provided similar results for common nodes, both frequently yielded higher node support relative to dataset A, arguing that: (i) more data yield increased node support and (ii) partial gene augmentation does not introduce an obvious nonphylogenetic bias. A comparison of single-gene bootstrap analyses identified four nodes for which one or two of the 25 genes provided modest to strong support for a grouping not recovered by the combined-gene result. As a summary proposal, two of these four groupings (one each within Bombycoidea and Lasiocampidae) were deemed sufficiently problematic to regard them as unresolved trichotomies. Since the alternative groupings were always highly localized on the tree, we did not judge a combined-gene analysis to present a problem outside those regions. Based on our robustly resolved results, we have revised the classification of Bombycoidea: the family Bombycidae is restricted to its nominate subfamily, and its tribe Epiini is elevated to subfamily rank (Epiinae stat.rev.), whereas the bombycid subfamily Phiditiinae is reinstated as a separate family (Phiditiidae stat.rev.). The bombycid subfamilies Oberthueriinae Kuznetzov & Stekolnikov, 1985, syn.nov. and Prismostictinae Forbes, 1955, syn.nov., and the family Mirinidae Kozlov, 1985, syn.nov. are established as subjective junior synonyms of Endromidae Boisduval, 1828. The family Anthelidae (Lasiocampoidea) is reincluded in the superfamily Bombycoidea.