Homopolish

Homopolish corrects systematic errors in nanopore sequencing assemblies to improve accuracy of microbial genome reconstruction.

Key Features:

• Systematic Error Correction: Uses homologous sequences for model training to identify and correct systematic nanopore sequencing errors that are often missed by methods trained on raw nanopore reads.
• Performance Superiority: Demonstrated superior error correction relative to Medaka and HELEN across datasets including bacteria, viruses, fungi, and metagenomic samples.
• Enhanced Genome Quality: When applied alongside or after Medaka or HELEN, can significantly improve assembly quality, reaching accuracies reported to exceed Q50 on R9.4 flow cells.

Scientific Applications:

• Microbial genome reconstruction: Enables high-accuracy reconstruction of bacterial, viral, and fungal genomes from nanopore-only sequencing data.
• Phylogenetic studies: Provides more accurate assemblies for sequence-based comparisons used in phylogenetic analyses.
• Functional annotation: Reduces sequencing-induced errors to improve the reliability of downstream functional annotation.
• Pathogen identification: Facilitates pathogen detection from nanopore data by producing higher-quality assemblies.
• Metagenomic analysis: Improves assembly accuracy in metagenomic samples to support community-level analyses.

Methodology:

Homopolish trains a model using homologous sequences and leverages evolutionary conservation among related sequences to identify and correct systematic errors specific to nanopore sequencing data.