Efficient error correction for next-generation sequencing of viral amplicons

Area of Science:

• Bioinformatics
• Genomics
• Virology

Background:

• Next-generation sequencing (NGS) enables analysis of numerous viral sequence variants, crucial for studying virus evolution, drug resistance, and immune escape.
• Bulk sequencing is prone to errors, necessitating robust error identification and correction methods.
• Existing error correction methods are often not optimized for amplicon analysis and assume random error distribution, failing to account for sequence-specific error patterns like homopolymers.

Purpose of the Study:

• To develop and evaluate novel, efficient error correction algorithms specifically optimized for viral amplicon sequencing data.
• To address the limitations of existing methods by incorporating sequence-specific error characteristics.

Main Methods:

• Development of two new algorithms: k-mer-based error correction (KEC) and empirical frequency threshold (ET).
• Comparison of KEC and ET against a previously published clustering algorithm (SHORAH).
• Performance evaluation using 24 experimental datasets of 454-sequencing amplicons with known sequences.

Main Results:

• KEC and ET demonstrated high accuracy in identifying true viral haplotypes, comparable to SHORAH.
• KEC and ET significantly outperformed SHORAH in removing false haplotypes.
• KEC and ET provided more accurate estimations of true haplotype frequencies.

Conclusions:

• The KEC and ET algorithms are highly effective for the rapid recovery of error-free haplotypes from heterogeneous viral amplicon sequencing data.
• These algorithms are particularly suitable for 454-sequencing data.
• Implementations and datasets are publicly available for use and further research.