CANU

CANU assembles de novo genomes from noisy single-molecule long reads generated by Pacific Biosciences (PacBio) and Oxford Nanopore to produce high-quality reference assemblies.

Key Features:

• Support for Long-Read Sequencing: Optimized for long-read data from Pacific Biosciences (PacBio) and Oxford Nanopore for improved reconstruction of reference genomes.
• Error Rate Management: Implements algorithmic strategies to handle high error rates in single-molecule long reads while resolving large repeats and closely related haplotypes.
• Adaptive Overlapping Strategy: Uses a tf-idf weighted MinHash adaptive overlapping strategy to improve overlap detection accuracy and assembly continuity.
• Sparse Assembly Graph Construction: Constructs sparse assembly graphs to prevent collapse of diverged repeats and haplotypes for more accurate assemblies.
• Reduced Depth-of-Coverage Requirements: Lowers depth-of-coverage requirements by approximately half compared to Celera Assembler 8.2.
• Improved Runtime Efficiency: Achieves substantial runtime reductions—by an order of magnitude for large genomes—relative to earlier versions.
• High Assembly Continuity and Quality: Capable of producing complete microbial genomes and near-complete eukaryotic chromosomes, achieving contig NG50 > 21 Mbp on human and Drosophila melanogaster PacBio datasets.

Scientific Applications:

• Complex genome assembly: Reconstruction of microbial and eukaryotic genomes from long-read sequencing data.
• Reference-quality genome generation: Automated production of reference-quality genome assemblies for downstream genomic analyses.
• Graph-based integration: Outputs assembly graphs in GFA format for integration with phasing and scaffolding techniques.

Methodology:

Uses novel overlapping and assembly algorithms including an adaptive overlapping strategy based on tf-idf weighted MinHash and sparse assembly graph construction to handle high error rates and avoid collapse of diverged repeats and haplotypes.