Accurate long-read de novo assembly evaluation with Inspector

Area of Science:

• Genomics
• Bioinformatics

Background:

• Long-read sequencing technologies have advanced de novo genome assembly.
• Accurate evaluation of genome assembly quality, particularly for structural variations, remains challenging.
• Existing tools often require a reference genome, limiting their applicability for de novo assembly assessment.

Purpose of the Study:

• To introduce Inspector, a novel reference-free tool for evaluating long-read de novo genome assemblies.
• To enable precise identification and localization of structural errors in genome assemblies.
• To provide a method for correcting assembly errors using raw sequencing reads.

Main Methods:

• Inspector utilizes raw long reads to perform reference-free evaluation of genome assemblies.
• The tool identifies various types of assembly errors, including large-scale and small-scale structural variations.
• Error correction is achieved by generating consensus sequences from raw reads that span erroneous regions.

Main Results:

• Inspector accurately identifies and reports the types and locations of assembly errors.
• The tool demonstrates effectiveness on both in silico datasets and real-world long-read assembly results.
• Inspector successfully corrects assembly errors, improving overall assembly quality.

Conclusions:

• Inspector offers a robust solution for the critical need of accurate, reference-free evaluation of long-read de novo genome assemblies.
• The tool's ability to identify and correct structural errors significantly enhances genome assembly quality assessment.
• Inspector represents a valuable advancement for the genomics community, facilitating more reliable genome reconstructions.