Skip to main content


Fig. 1 | Mobile DNA

Fig. 1

From: A call for benchmarking transposable element annotation methods

Fig. 1

Variation among TE annotation tools. a TE coverage in the Arabidopsis thaliana genome resulting from three commonly used repetitiveness-based de novo tools, compared to a reference set of TEs [8]. The total amount of TE coverage differs between the three, as does the fraction of the reference TEs that were found or missed and the amount of non-reference putative TEs. b Full-length LTR TEs in the Drosophila melanogaster X chromosome found by five different LTR-specific de novo tools, compared to a reference set of TEs [24]. Similar to a but even more pronounced, the number of TEs found by the tools and their agreement with the reference set vary widely. c A 100-kbp segment of the Arabidopsis lyrata genome (scaffold_1:14,957,501-15,057,500) displayed on a custom UCSC genome browser [76, 77], illustrating differences among TE annotations resulting from several approaches, as well as additional genomic data useful in identifying bona fide TEs. From top to bottom, the tracks represent: RepeatMasker annotations using libraries from Repbase [37], RepeatModeler [30], REPET [44], or de la Chaux et al. [78]; full-length LTR TE predictions by LTR_Finder [33] or LTRharvest [79]; tandem repeat predictions by TRF [29]; gene models predictions by FGenesH [80]; a set of TE-specific domains [13]; mapped mRNA and small RNA short reads [77]; inter-species conservation (alignment percent identity plots) to other Brassicaceae species [77]; and genome self-alignment depth (generated with LASTZ)

Back to article page