2.4 Transcriptome sequencing, assembly and annotation
Libraries were sequenced as 150 bp, paired-end (PE) reads using an Illumina HiSeq4000 sequencing system, performed by Novogene (Hong Kong, China). Raw data have been deposited in the NCBI Short Read Archive under the accessions numbers (SRA: SRR10426055-76). Adapter sequences were trimmed and low quality reads were filtered using a 5 bp sliding window with a mean quality threshold of 20 and minimum read length of 36 bp using Trimmomatic v0.36 (Bolger, Lohse, & Usadel, 2011). Read quality (before and after quality-filtering) was assessed using FastQC v0.11.5 (Andrews, 2010).
Processed reads were assembled de novo  with Trinity v.2.5.1 (Grabherr et al., 2011). Separate reference transcriptomes were created for B. calyciflorus s.s. and B. fernandoi , using all reads generated for the respective species. To filter possible contamination, we used a custom perl script which, using the blastn algorithm (ncbi-blast-2.6.0; Camacho et al., 2015), assigns all contigs either to a local algae database (Monoraphidium minutum ,Chlamydomonas reinhardtii and, Cryptomonas sp) or to the respective B. calyciflorus s.s. genome assembly (Kim et al., 2018). Transcripts were only assigned as of rotifer origin when the top hit was to the B. calyciflorus s.s. genome and the bit-score gain over matches to the next species was >100. The same custom script was further used to remove ribosomal RNA reads by performing a blastn search to a local database consisting of 18S and 28S sequences ofBrachionus species downloaded from NCBI. We did not utilize theB. calyciflorus s.s. reference genome for the analyses presented here, as this would introduce a bias, since the B. fernandoireads would not map as well to the reference as those from B. calyciflorus s.s..