Fig. 1

Schematic representation of RM2 output. A Sequences that are part of a LTR element. The internal and 5′/3′ LTR “subparts” are often presented as two different fasta entries / prospective families to allow for solo-LTRs (see solo-LTR in Other TEs). In the case of RepeatModeler2, the internal subpart is labelled “type = INT” (blue) and the LTR “Type = LTR” (red). In some cases “Type = LTR” can also refer to full-length LTR elements (combined red-blue-red blocks, as presented in the figure). RM2 will attempt to classify the family and this information can be found after the “#” symbol. In this example, the LTR subpart (red) is classified as Order LTR, superfamily Unknown and the internal subpart (blue) is classified also as Order LTR but superfamily “Gypsy”. When the LTR subpart and internal are reported in separate consensus sequences, it is possible to use manual curation to find the ones that form one full-length LTR element (see main text). The number after “Final Multiple Alignment Size” indicates how many sequences were used in the multiple sequence alignment to generate this consensus. B Sequence representing a prospective PLE predicted by Repeatmodeler2. The fasta header includes the classification, the software used for the prediction (RepeatScout), prospective family size and potential family relationships. Notice that the Family size (number of copies found in the genome) is different from the Final Multiple Alignment Size, which is limited to 100 to ease computation. Also note that PLEs are not LINEs (see Appendix I in Supplementary Methods), although for historic reasons they are classified as such by several databases and pipelines