LTR_retriever

LTR_retriever identifies long terminal repeat retrotransposons (LTR-RTs) de novo from genomic sequences, producing high-quality LTR libraries to support genome annotation and evolutionary analyses.

Key Features:

• High sensitivity and specificity: Achieves reported performance metrics of 91% sensitivity, 97% specificity, 96% accuracy, and 90% precision in the rice (Oryza sativa) genome.
• Multithreading: Implements multithreaded processing to increase throughput on large genomic datasets.
• Compatibility with long sequencing reads: Supports long-read data and was validated using 40k self-corrected PacBio reads (~4.5× coverage) in Arabidopsis thaliana while maintaining high sensitivity and specificity.
• Identification of noncanonical LTR-RTs: Detects non-TGCA LTR termini in addition to canonical 5'-TG…CA-3' termini and identified seven types of noncanonical LTRs across 42 of 50 surveyed plant genomes, predominantly among Copia elements.
• Generation of LTR libraries: Produces high-quality LTR libraries for downstream annotation and analysis.
• Insights into insertion preferences: Detects non-TGCA Copia elements that frequently reside in or near genes, indicating potential roles in gene evolution and mutagenesis studies.

Scientific Applications:

• Genome annotation: Provides comprehensive identification of canonical and noncanonical LTR-RTs to improve repeat masking and gene annotation in plant genomes.
• Evolutionary analyses: Enables studies of retrotransposon impact on gene evolution by characterizing insertion patterns and LTR-RT diversity.
• Long-read assembly analysis: Facilitates LTR-RT detection from long-read sequencing datasets such as self-corrected PacBio reads.

Methodology:

Performs structure-based de novo identification of LTR-RTs from genomic sequences by leveraging conserved LTR-RT structural features and generates high-quality LTR libraries for downstream analyses.