Evolutionary insights from comparative transcriptome and transcriptome-wide coalescence analyses in Tetrastigma hemsleyanum

Yihan Wang, Weimei Jiang, Wenqing Ye, Chengxin Fu, Matthew A Gitzendanner, Pamela S Soltis, Douglas E Soltis, Yingxiong Qiu

Tetrastigma hemsleyanum is of great medicinal importance and used as a model system to address the evolutionary history of warm-temperate evergreen (WTE) forest biomes in East Asia over Neogene time scales. However, further studies on the neutral and adaptive divergence processes of T. hemsleyanum are currently impeded by a lack of genomic resources. In this study, we de novo assembled and annotated a reference transcriptome for two cpDNA lineages (Central-South-East vs. Southwest) of T. hemsleyanum. We further used comparative genomic and multilocus coalescent approaches to investigate the tempo and mode of lineage diversification in T. hemsleyanum. A total of 52,838 and 65,197 unigenes with an N50 of 1,667 and 1,841 bp for Central-South-East (CSE) and Southwest (SW) lineages, respectively, were recovered, and 6,692 putative orthologs were identified between the two lineages. Estimation of Ka/Ks ratios for these orthologs revealed that ten genes had Ka/Ks values significantly greater than 0.5 (P < 0.05), whereas 2,099 (Ka/Ks < 0.5, P < 0.05) were inferred to be under purifying selection. Based on three bioinformatic strategies, we identified a total of 1,018 single-copy nuclear genes (SCNGs) from the orthologs. We successfully designed eight nuclear gene primer pairs with high intraspecific variation (e.g. hT = 0.923, πT = 1.68×10-3), when surveyed across a subset of T. hemsleyanum individuals. Concordant with the previous cpDNA data, the haplotype networks constructed for most nuclear gene loci clearly identified the two lineages. A multilocus coalescence analysis suggested that the separation between the two lineages appears to have occurred during the mid-Pliocene. Despite their ancient divergence, both lineages experienced expansion at rather localized scales and have continued to exchange genes at a low rate. This study demonstrated the utility of transcriptome sequencing as a basis for SCNG development in non-model species and the advantages of integrating multiple nuclear loci for phylogeographic and phylogenetic studies.