Ratatosk

Ratatosk corrects base-level errors in Oxford Nanopore long reads by leveraging accurate short reads from short-read sequencing (SRS) to reduce long-read sequencing (LRS) error rates and improve sequence accuracy for downstream genomic analyses.

Key Features:

• Phased Hybrid Error Correction: Reduces high error rates of Oxford Nanopore long reads by combining long reads from LRS and accurate short reads from SRS.
• Compacted and Colored de Bruijn Graphs: Builds compacted, colored de Bruijn graphs from short reads to identify correct sequences for error correction of long reads.
• fc-mer Matching: Uses fc-mer matches to anchor long reads onto the de Bruijn graph to guide sequence correction.

Scientific Applications:

• Improved Variant Calling: Produces corrected long reads that enable nearly 99% SNP call accuracy and up to ~40% improvement in indel call accuracy.

Methodology:

Constructs compacted, colored de Bruijn graphs from accurate short reads; anchors Oxford Nanopore long reads onto the graph using fc-mer matches; and performs phased hybrid error correction by integrating short-read information within the graph.