misFinder: identify mis-assemblies in an unbiased manner using reference and paired-end reads

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing generates short reads, leading to genome assembly errors that impact downstream analysis.
• Existing error correction tools have limitations, including inability to distinguish real structural variations or producing false positives.

Purpose of the Study:

• To develop misFinder, a tool for accurate and unbiased identification and correction of genome assembly errors.
• To enhance genome assembly accuracy for improved downstream data analysis.

Main Methods:

• Combines reference genome information with aligned paired-end reads to the assembled sequence.
• Detects assembly errors and structural variations by comparing reference and assembled genomes.
• Analyzes aligned paired-end reads using coverage and insert distance features to distinguish error types and ensure high-confidence calls.

Main Results:

• misFinder accurately identifies assembly errors with high confidence and minimal miscalls.
• The tool effectively distinguishes correct assemblies corresponding to structural variations from mis-assembled sequences.
• Demonstrated superior performance over QUAST and REAPR in identifying true positive mis-assemblies with reduced false positives/negatives.

Conclusions:

• misFinder significantly improves genome assembly accuracy.
• The tool offers a robust solution for identifying and correcting assembly errors, outperforming existing methods.
• misFinder is available for free download, facilitating its use in genomic research.