RepeatModeler2

RepeatModeler2 identifies and models transposable element (TE) families de novo from genome sequences to enable discovery and annotation of TEs.

Key Features:

• Integrated de novo repeat discovery: Integrates RECON, RepeatScout, and LtrHarvest/Ltr_retriever to identify repeat element boundaries and infer family relationships from sequence data.
• Structural LTR discovery: Includes a module for the structural discovery of complete long terminal repeat (LTR) retroelements.
• Benchmark performance: Benchmarked on Drosophila melanogaster, Danio rerio, and Oryza sativa, identifying approximately three times more consensus sequences matching manually curated sequences at >95% sequence identity and coverage, with particular improvement for LTR retroelements.
• Low false positive rate: Demonstrated an extremely low false positive rate when applied to simulated genomes devoid of TEs.

Scientific Applications:

• Unsupervised TE annotation: Generates de novo TE consensus sequences and family models for annotation of genome assemblies.
• Evolutionary genomics: Supports study of the composition and impact of parasitic, self-mobilizing TEs on organismal evolution.
• Reference library generation: Produces comprehensive reference libraries that reflect the composition of complex TE landscapes.

Methodology:

Combines RECON, RepeatScout, and LtrHarvest/Ltr_retriever to de novo compile consensus sequence models representing unique TE families dispersed in the genome and to perform structural discovery of LTR retroelements.