Genome size estimation and RADseq library preparation
For three species (Adelobotrys adscendens , Axinaea costaricensis , Meriania phlomoides ), we estimated genome sizes to select appropriate restriction enzymes for RADseq. We prepared fresh leaf material for three individuals per species in Otto’s buffer (Otto et al. 1981) for propidium iodide flow cytometry (CyFlow ML flow cytometer, Partec, Münster, Germany; 532nm/100mW laser, Cobolt Samba, Cobolt AB, Solna, Sweden) following Temsch et al. (2010). We used Solanum pseudocapsicum (1C = 1.29 pg DNA) as internal standard to calculate 1C values for each run (three runs per sample) and calculated the mean 1C per species over all runs and individuals. Average genome sizes (1C value) for the three species were estimated as follows: Ad. adscendens 0.344 pg, sd 0.003; A. costaricensis 0.587 pg, sd 0.01; M. phlomoides 0.639 pg, sd 0.006. According to the Kew C-value database, these genome sizes are the second, third and fourth estimates for the family (Hanson et al. 2001).
In accordance with the relatively small genome sizes, we selected the restriction enzyme PstI (New England Biolabs, Ipswich, MA, USA) for single-digest RADseq (protocol modified from Paun et al. (2016)). We prepared eight RADseq libraries starting with 240 ng DNA per sample, pooling 72 samples per library, including also some repeats (using 120 ng DNA per sample). We ligated 300 nM P1 barcoded adapters (150 nM if 120 ng input DNA) to the restricted samples at 16°C overnight. The P1 adapters included both index and inline barcodes, that were different from each other by at least 3 nucleotide positions. After P1 ligation, we sheared the DNA by sonication in a Bioruptor Pico (Diagenode, Seraing, Belgium) using two cycles (45 seconds ‘on’, 30 seconds ‘off’; at 4°C) to obtain DNA fragments of 400 bp on average. After P2 adapter ligation, PCR amplification and clean-up steps (using the MiniElute PCR purification Kit, Qiagen), we performed a final size selection for the range 220 to 850 bp using a 1.5% precasted dye-free cassette and a Pippin Prep (Sage Science). All libraries were sequenced on an Illumina HiSeq 2500 machine at Vienna BioCenter Core Facilities (VBCF) (http://www.viennabiocenter.org/vbcf/next-generation-sequencing/) as 100 bp single-end reads.