Fig. 2

Overview of the process of manual curation. Graphical representation of the steps taken during the manual curation of a putative TE family. A A query sequence is selected to start the process. The selection can be based on the priorities that make it more likely to be an autonomous TE. B The putative TE is used as a query in a blast search against the reference genome and hits are recorded. C In most cases, the prospective families are a truncated version of the true TE, and thus the hit coordinates in the genome will also be truncated. To capture as much TE sequence as possible from the genomic location, the hit coordinates are extended by a given number of bases up- and down-stream in the genome. D The genomic sequences are extracted from the genome using ‘bedtools getfasta’, a multiple sequence alignment (MSA) is generated and written to a file. E The MSA is visualised in an alignment viewer and manually curated and a consensus is generated. F The final curated TE consensus is compared to itself to check for completeness. Using a dot-plot is particularly diagnostic in the case of elements flanked by repeats (as in the case of LTRs) but it is also possible to assess completeness by searching for conserved protein domains