Parsing TE Superfamily Amplification

To summarize the overall amplification history of TE superfamilies and test for ongoing activity, the perl script parseRM.pl (Kapusta et al., 2017 (link)) was used to parse the raw output files from RepeatMasker (.align) and report the sequence divergence between each read and its respective consensus sequence (parameter values = -l 50,1 and -a 5). The repeat library used to mask the reads comprised the 55,327 TE contigs classified by the PiRATE pipeline and clustered at 100% sequence identity. Each TE superfamily is therefore represented by multiple consensus sequences corresponding to the family and subfamily TE taxonomic levels (i.e., not the distant common ancestor of the entire superfamily). For each superfamily, histograms were plotted to summarize the percent divergence of all reads from their closest (i.e., least divergent) consensus sequence. These histograms do not allow the delineation between different amplification dynamics scenarios (i.e., a single family with continuous activity versus multiple families with successive bursts of activity). Rather, these global overviews were examined for overall shapes consistent with ongoing activity (i.e., the presence of TE loci <1% diverged from the ancestral sequence and a unimodal, right-skewed, J-shaped, or monotonically decreasing distribution).