Skip to main content

Table 1 Benchmarking whole-genome TE annotations on plant and animal genomes including chicken, fly, maize, mouse, rice, zebrafish, and zebra finch

From: Response to Commentary: Accounting for diverse transposable element landscapes is key to developing and evaluating accurate de novo annotation strategies

Category

TE

Benchmark

EDTAa

EDTA_repbaseb

EDTA_repeatmodeler2c

RepeatModeler2d

Animals

LTR

Sensitivity

63.3%

57.6%

72.2%

81.4%

Animals

non-LTR

Sensitivity

1.0%

43.7%

66.9%

76.1%

Animals

TIR

Sensitivity

38.8%

35.1%

40.9%

38.9%

Plants

LTR

Sensitivity

94.7%

94.6%

94.7%

73.5%

Plants

non-LTR

Sensitivity

0.0%

82.2%

49.6%

57.5%

Plants

TIR

Sensitivity

67.1%

66.8%

67.1%

28.1%

Animals

LTR

Specificity

95.0%

97.8%

96.9%

96.8%

Animals

non-LTR

Specificity

100.0%

98.2%

98.7%

97.9%

Animals

TIR

Specificity

88.4%

91.7%

89.8%

97.4%

Plants

LTR

Specificity

90.4%

90.7%

90.5%

97.9%

Plants

non-LTR

Specificity

100.0%

99.6%

99.9%

99.9%

Plants

TIR

Specificity

94.7%

95.1%

94.8%

99.6%

  1. aEDTA run with the --sensitive 1 parameter
  2. bEDTA run with the --curatedlib repbase-nonLTR.fasta input
  3. cEDTA run with the --curatedlib RepeatModeler2-nonLTR.fasta input
  4. dRepeatModeler2 run with default parameters