BCOOL

BCOOL corrects short-read sequencing errors using compacted de Bruijn graphs to improve data accuracy for genome assembly and hybrid long-read correction.

Key Features:

• De Bruijn Graph-Based Correction: Employs compacted de Bruijn graphs to leverage full-length read information for error correction rather than treating reads only as k-mer spectra.
• Graph Construction and Filtering: Constructs a compacted de Bruijn graph from input reads and filters it by k-mer abundance and unitig abundance to remove sequencing errors.
• Read Mapping for Correction: Maps original reads to the cleaned graph and corrects them, yielding higher accuracy than traditional k-mer-spectrum correctors.
• Scalability: Scales to human-size genomic datasets and beyond for large-scale sequencing projects.

Scientific Applications:

• Short-read error correction: Improves accuracy of short-read sequencing data for downstream analyses.
• Genome assembly: Produces corrected reads to enhance de novo genome assembly quality.
• Hybrid long-read correction: Provides short-read-based corrections to support hybrid correction of long-read sequencing data.

Methodology:

Constructs a compacted de Bruijn graph from input short reads; filters the graph by k-mer abundance and unitig abundance; maps reads to the cleaned graph to perform corrections; implemented using components from Bcalm2, Bgreat2, Btrim, and Ntcard.