Figure 1.
Conflicting phylogenetic relationships among extant gymnosperms.
(A) The four main lineages of extant gymnosperms: (1) conifers (Pinus resinosa), (2) cycads (Cycas sp.), (3) Ginkgo biloba, and (4) gnetophytes (Ephedra chilensis). (B) Two main hypotheses for phylogenetic relationships of gymnosperms. (C) Three main hypotheses for the phylogenetic placement of gnetophytes. (D) Two main hypotheses for the phylogenetic placement of Ginkgo.
Figure 2.
Species trees inferred from (A) 305 nuclear genes and (B) 47 plastid genes using the coalescent method (STAR).
Bootstrap percentages (BPs) from STAR/RAxML are indicated above each branch; an asterisk indicates that the clade is supported by 100 BPs from both STAR and RAxML. Branch lengths were estimated by fitting the concatenated matrices to the inferred topology from STAR.
Figure 3.
Summary of bootstrap percentages (BPs) from coalescent and concatenation analyses using different gene subsampling and rate partitions.
(A) BPs from coalescent and concatenation analyses using different gene subsampling. The 305 nuclear genes were subsampled for four different gene size categories (i.e., 25, 47, 100, or 200 genes; 10 replicates each), and the 47 plastid genes were subsampled for 25 genes (10 replicates). Cells with hatching indicate that support for the placement of Ginkgo biloba from all replicates is below 80 BP; colored cells indicate relationships that received bootstrap support ≥80 BP from at least one replicate (pink = cycads plus Ginkgo as sister to all remaining extant gymnosperms, yellow = Ginkgo alone as sister to conifers and gnetophytes within extant gymnosperms; see also Figure 1D). (B) BPs from coalescent and concatenation analyses across different nucleotide rate partitions. Parsimony informative sites in concatenated matrices were sorted based on estimated evolutionary rates, and subsequently divided into two equal partitions. The index of substitution saturation (ISS) was used to measure nucleotide substitution saturation for sites within each rate partition. The two critical ISS values, i.e., ISS.C1 and ISS.C2, were estimated using an asymmetrical and symmetrical topology, respectively (for data including more than 32 species, only values estimated from 32 terminals are shown here).