Comparative plastome analysis of Musaceae and new insights into phylogenetic relationships

Abstract

Background: Musaceae is an economically important family consisting of 70-80 species. Elucidation of the interspecific relationships of this family is essential for a more efficient conservation and utilization of genetic resources for banana improvement. However, the scarcity of herbarium specimens and quality molecular markers have limited our understanding of the phylogenetic relationships in wild species of Musaceae. Aiming at improving the phylogenetic resolution of Musaceae, we analyzed a comprehensive set of 49 plastomes for 48 species/subspecies representing all three genera of this family.

Results: Musaceae plastomes have a relatively well-conserved genomic size and gene content, with a full length ranging from 166,782 bp to 172,514 bp. Variations in the IR borders were found to show phylogenetic signals to a certain extent in Musa. Codon usage bias analysis showed different preferences for the same codon between species and three genera and a common preference for A/T-ending codons. Among the two genes detected under positive selection (dN/dS > 1), ycf2 was indicated under an intensive positive selection. The divergent hotspot analysis allowed the identification of four regions (ndhF-trnL, ndhF, matK-rps16, and accD) as specific DNA barcodes for Musaceae species. Bayesian and maximum likelihood phylogenetic analyses using full plastome resulted in nearly identical tree topologies with highly supported relationships between species. The monospecies genus Musella is sister to Ensete, and the genus Musa was divided into two large clades, which corresponded well to the basic number of n = x = 11 and n = x =10/9/7, respectively. Four subclades were divided within the genus Musa. A dating analysis covering the whole Zingiberales indicated that the divergence of Musaceae family originated in the Palaeocene (59.19 Ma), and the genus Musa diverged into two clades in the Eocene (50.70 Ma) and then started to diversify from the late Oligocene (29.92 Ma) to the late Miocene. Two lineages (Rhodochlamys and Australimusa) radiated recently in the Pliocene /Pleistocene periods.

Conclusions: The plastome sequences performed well in resolving the phylogenetic relationships of Musaceae and generated new insights into its evolution. Plastome sequences provided valuable resources for population genetics and phylogenetics at lower taxon.

Keywords: Divergence time; Musaceae; Phylogeny; Plastome; cpDNA marker.

Fig. 1

Distribution maps for (A) three genera of Musaceae and (B) five sections of genus Musa

Fig. 2

Plastome map of Musaceae species generated in this study. Genes inside the outer circle are transcribed clockwise while those outsides are transcribed counterclockwise. Genes are color-coded according to their function. Darker gray columns in the inner circle represent the GC content and the lighter gray columns accordingly correspond to the AT content. Genes marked with one asterisk (*) contain at least one intron

Fig. 3

Codon content of 20 amino acids and stop codon in Musaceae plastomes. The relative synonymous codon usage (RSCU) values are shown on the y-axis

Fig. 4

Analysis of repeat elements in Musaceae plastomes. A Frequency and average proportion of four types of short dispersed repeats (SDRs). Pie chart showing the average proportion of four SDRs types. B Frequency and average proportion of SDRs by length. C Frequency and average proportion of six simple sequence repeats (SSRs) types. D Frequency and average proportion of SSRs by base composition. E Average proportion of SSRs in coding and noncoding region. F Average proportion of SSRs in the LSC, SSC, and IR. The phylogenetic tree was infered from comlete plastome dataset

Fig. 5

Maximum likelihood (ML) cladogram (A) and phylogram (B) of Musaceae inferred from complete plastomes using RAxML. ML bootstrap (BS) values and the posterior probabilities (PP) calculated from MrBayes are shown at nodes, except nodes with 100% BS and 1.0 PP, ‘-’indicates PP under 0.5. Clade is set to polytomy when BS<50% and PP<0.5

Fig. 6

Divergence time of Musaceae obtained from BEAST 2 based on the five genes (ccsA, matK, ndhF, rpoC1, and rpoC2) selected by Sortadate. Mean divergence times of the nodes are shown at the nodes and the blue bars correspond to the 95% highest posterior density (HPD). The outgroup taxa are marked with gray strips