Benchmarking methods for genome annotation using nanopore direct RNA in a non-model crop plant

Area of Science:

• Plant genomics
• Bioinformatics
• Molecular biology

Background:

• High-quality genome annotations are crucial for understanding plant responses to environmental stressors.
• Direct RNA sequencing with nanopore long reads is a powerful method for improving genome annotations.
• Existing tool benchmarking studies predominantly focus on animal models, necessitating plant-specific evaluations.

Purpose of the Study:

• To benchmark five genome annotation tools (StringTie3, IsoQuant, Bambu, FLAIR, FLAMES) using direct RNA sequencing data from barley.
• To assess tool performance in isoform detection, structural completeness, splicing classification, and handling of 5' read truncation.
• To identify optimal tools for improving genome annotations in non-model plant species.

Main Methods:

• Direct RNA sequencing of barley infected with Net Form Net Blotch disease.
• Benchmarking of five annotation tools: StringTie3, IsoQuant, Bambu, FLAIR, and FLAMES.
• Comparative analysis of tool performance based on transcript novelty, accuracy, and completeness.

Main Results:

• Substantial variation in performance was observed across the five annotation tools.
• The top-performing reference-guided tools identified over 700 novel transcripts.
• Identified novel transcripts include candidates with potential roles in plant disease response.
• Demonstrated the utility of direct RNA sequencing for enhancing barley genome annotation.

Conclusions:

• Plant-specific benchmarking of bioinformatic tools is essential for accurate genomic analysis.
• Direct RNA sequencing is highly effective for improving genome annotations in plant species.
• The study provides valuable insights for enhancing reference genomes of non-model plants, particularly for stress response studies.