Nymphoides peltata was planted in the greenhouse at the Wuhan Botanical Garden of the Chinese Academy of Sciences, China. The floral bud samples from four developmental stages (i.e., 1, 3, 5, and 8 mm) of the same individual for each morph were collected and stored subsequently at −80°C until RNA extraction.

For each developmental stage of each morph, total RNA was extracted using RNAiso Plus reagent (TaKaRa, Dalian, China) and was treated with RNase-free DNase I (TaKaRa, Dalian, China). Quality and integrity of the extracted total RNA was determined by using Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Concentrations of the total RNA were measured using Qubit® 2.0 Flurometer (Life Technologies, CA, USA). Equal amounts of total RNA (7 μ g) from each stage of long- or short-style flowers were pooled together, respectively. These two pooled RNA samples were sent to BGI (Beijing Genomic Institute) for RNA sequencing (BGI Inc., Shenzhen, China). Two cDNA libraries (referred as L1 and S1) from these two RNA samples were prepared using Illumina TruSeqRNA sample Preparation Kit (Illumina), and were sequenced using the Illumina Hiseq2000 (Illumina Inc., San Diego CA, USA) to generate 100 bp paired-end raw reads.

The raw reads were trimmed using the program Trimmomatic v0.32 (Bolger et al., 2014) by removing adapter sequences, sequences with unknown base call (N) more than 5%, and low quality sequences (<Q20). Transcriptome De novo assembly was carried out via a short read assembling program, Trinity (version trnityrnaseq_r2012-10-05) (Grabherr et al., 2011), with min_kmer_cov set to 2 by default and all other parameters set as their defaults. The statistics of the transcriptome assemblies are shown in Table 1.

To improve efficiency in searching orthologs between long- and short-styled flowered plants transcriptomes of N. peltata, the genome sequence of Salvia miltiorrhiza (ftp://202.203.187.112/genome/danshen/) was chose as reference because among the plant species whose whole genome had been sequenced, the S. miltiorrhiza was shown to be more closely related to N. peltata. Identification of orthologous groups among these three samples was firstly assessed by using the program, OrthoMCL v1.4 (Li et al., 2003), using default settings. Those genes with lengths longer than 50 amino acids and orthologous groups with a single copy gene for each of the three samples were used for further analysis. Then, the sequences of S. miltiorrhiza were removed and each remaining orthologous group was aligned using the program MUSCLE v3.8.31 (Edgar, 2004) with default parameters. The poorly aligned regions, such as alignment length <200 bp or including unexpected stop codon, were excluded and the remaining sequences were used to generate maximum likelihood (ML) tree using the program PAML v4.8 (Yang, 2007) with the general time reversible (GTR) model.

The ratio of non-synonymous (Ka) and synonymous (Ks) substitution sites between pairs of orthologous genes identified from long- and short-styled flowered plants of N. peltata were estimated using a maximum likelihood (ML) method. Ka/Ks calculation was performed with PAML v4.8 (Yang, 2007) using default settings. High Ka/Ks (>>1) ratio indicated divergent orthologous genes and the positive selection may act on the orthologous gene, while low Ka/Ks (< <0.1) ratio indicated conserved orthologous genes. These divergent orthologous gene sequences can be used for further study on the genetic basis of distyly and adaptive evolution of distylous plant.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.