Data from: Phylogenetics of flowering plants based on combined analysis of plastid atpB and rbcL gene sequences

Following (1) the large scale molecular phylogeny of seed plants based on
plastid rbcL gene sequences (published in 1993 by Chase et al., Ann.
Missouri Bot. Gard. 80: 528-580) and (2) the 18S nuclear phylogeny of
flowering plants (published in 1997 by Soltis et al., Ann. Missouri Bot.
Gard. 84: 1-49), we present a phylogenetic analysis of flowering plants
based upon a second plastid gene, atpB, analyzed separately and in
combination with rbcL sequences for 357 taxa. Despite some discrepancies,
the atpB-based phylogenetic trees were highly congruent with those derived
from the analysis of rbcL and 18S rDNA, and the combination of atpB and
rbcL DNA sequences (comprising ca. 3000 base pairs) produced increased
bootstrap support for many major sets of taxa. The angiosperms are divided
into two major groups: noneudicots with inaperturate or uniaperturate
pollen (monocots plus Laurales, Magnoliales, Piperales, Ceratophyllales,
and Amborellaceae-Nymphaeaceae-Illiciaceae) and the eudicots with
triaperturate pollen (particularly asterids and rosids). Based on rbcL
alone and atpB/rbcL combined, the noneudicots (excluding Ceratophyllum)
are monophyletic, whereas they form a grade in the atpB trees.
Ceratophyllum is sister to the rest of angiosperms whith rbcL alone and in
the combined atpB/rbcL analysis, whereas with atpB alone, Amborellaceae,
Nymphaeaceae, and Illiciaceae/Schisandraceae form a grade at the base of
the angiosperms. The phylogenetic information at each codon position and
the different types of substitutions (observed transitions and
transversions in the trees versus pairwise comparisons) were examined;
taking into account their respective consistency and retention indices, we
demonstrate that third codon positions and transitions are the most useful
characters in these phylogenetic reconstructions. This study further
demonstrates that phylogenetic analysis of large matrices is feasible.

本研究基于两项前期研究成果：（1）Chase等人于1993年发表于《Ann. Missouri Bot. Gard.》第80卷528-580页的、基于质体rbcL基因（plastid rbcL gene）序列构建的大型种子植物分子系统发育树，以及（2）Soltis等人于1997年发表于《Ann. Missouri Bot. Gard.》第84卷1-49页的、被子植物18S核基因系统发育树。针对357个类群，我们选取第二质体基因atpB开展系统发育分析，并分别将其与rbcL序列单独或联合进行分析。
尽管存在部分分歧，但基于atpB的系统发育树与基于rbcL和18S rDNA的分析结果高度一致；联合atpB与rbcL的DNA序列（总长约3000个碱基对）可提升多数主要类群的自举支持率（bootstrap support）。
被子植物可划分为两大类群：一类为具无萌发孔或单萌发孔花粉的非真双子叶植物（涵盖单子叶植物以及樟目、木兰目、胡椒目、金鱼藻目，与Amborellaceae（无油樟科）、Nymphaeaceae（睡莲科）、Illiciaceae（八角科）类群），另一类为具三萌发孔花粉的真双子叶植物（尤以菊类与蔷薇类为代表）。
基于单独rbcL序列以及atpB与rbcL联合序列的分析显示，排除Ceratophyllum（金鱼藻属）后的非真双子叶植物为单系类群，而在单独以atpB构建的系统发育树中，该类群呈现为演化级。
在单独使用rbcL序列以及联合atpB与rbcL的分析中，Ceratophyllum（金鱼藻属）为其余被子植物的姊妹群；而在单独使用atpB的分析中，Amborellaceae（无油樟科）、Nymphaeaceae（睡莲科）以及Illiciaceae（八角科）/Schisandraceae（五味子科）类群构成被子植物基部的演化级。
本研究针对每个密码子位置（codon position）的系统发育信息，以及不同类型的碱基替换（即系统发育树中观测到的转换（transitions）与颠换（transversions），并与两两序列比对结果进行对比）展开分析；结合各自的一致性指数（consistency index）与保留指数（retention index），我们证实密码子第三位与碱基转换是本次系统发育重建中最具价值的特征。
本研究进一步证明，针对大型序列矩阵开展系统发育分析具备可行性。