Repeat and haplotype aware error correction in nanopore sequencing reads with DeChat.

Area of Science:

• Genomics
• Bioinformatics

Background:

• Accurate analysis of long-read sequencing data relies on effective error correction.
• Existing methods are often suboptimal for Nanopore R10 simplex reads (error rates <2%).

Purpose of the Study:

• To develop a novel error correction method, DeChat, specifically for Nanopore R10 simplex reads.
• To enable repeat- and haplotype-aware error correction without overcorrection.

Main Methods:

• DeChat integrates de Bruijn graphs with variant-aware multiple sequence alignment.
• This synergistic approach preserves variants in repeats and haplotypes while correcting sequencing errors.

Main Results:

• DeChat reduces read errors by up to two orders of magnitude compared to existing methods.
• Benchmarking on simulated and real datasets confirms significant error reduction without data loss.

Conclusions:

• DeChat provides highly accurate error correction for Nanopore R10 simplex reads.
• DeChat-corrected reads enhance downstream applications like genome assembly and taxonomic classification.