Fig. 1

Overview of the PiRATE pipeline. Step 0: genome assembly and raw Illumina data are used as input data. Step 1: The detection of putative TEs and repeated sequences is performed using 12 tools, combining four detection approaches. Detected sequences from approaches 1 and 4 are filtered according to their length (minimum 500 bp). Detected sequences from the tools MITE-Hunter and SINE-Finder are directly saved as non-autonomous TEs. Other detected sequences are clustered with CD-HIT-est to reduce redundancy. Step 2: Putative TE sequences are automatically classified with PASTEC as potentially autonomous TEs, non-autonomous TEs or uncategorized sequences. The potentially autonomous TEs are manually checked and grouped into TE families. Step 3: Three libraries are manually constructed with a “Russian doll” strategy: 1) a “potentially autonomous TEs library”, a “total TEs library” and a “repeated elements library”. A double-run of TEannot is carried out for each library to select sequences that align with a full-length (FLC) on the genome assembly and finally obtain three independent annotations