ORNA

ORNA performs read normalization of high-coverage sequencing datasets to reduce redundancy while preserving k-mers and their relative abundances to maintain de Bruijn graph connectivity for genome and transcriptome assembly.

Key Features:

• Set Multi-Cover Optimization: Formulates normalization as a set multi-cover problem that minimizes the number of reads required while ensuring retention of original k-mers and their relative abundances and preserving de Bruijn graph connections.
• Preservation of K-Mer Abundances: Guarantees retention of all k-mers and their relative abundances from the original dataset.
• Performance and Efficiency: Reduces dataset size and computational resource requirements with reported assembly quality loss between 1% and 30% depending on reduction stringency.
• Error-Corrected Data Application: Normalizing error-corrected reads yields more accurate assemblies than normalizing uncorrected datasets.
• Novel Transcript Prediction: Combining and normalizing multiple RNA-seq datasets enables prediction of novel transcripts that may be missed otherwise.
• General-Purpose Normalization: Applicable across diverse RNA-seq datasets with varying coverage to reduce redundancy while maintaining assembly-relevant information.

Scientific Applications:

• Genome assembly: Supports de Bruijn graph-based genome assembly by preserving k-mers and graph connectivity during read reduction.
• Transcriptome assembly: Supports transcript reconstruction by maintaining k-mer abundances critical for assembling transcripts from RNA-seq data.
• Novel transcript discovery: Enables detection of previously unidentified transcripts by integrating and normalizing multiple datasets.
• High-coverage data management: Reduces redundancy and computational resource demands in large-scale sequencing projects.
• Preprocessing for error-corrected data: Enhances assembly accuracy when applied to error-corrected sequencing reads.

Methodology:

ORNA formulates read normalization as a set multi-cover optimization that selects a minimal subset of reads while guaranteeing retention of all k-mers and their relative abundances to preserve de Bruijn graph connectivity.