Comprehensive evaluation and guidance of structural variation detection tools in chicken whole genome sequence data

Area of Science:

• Genomics
• Bioinformatics
• Comparative Genomics

Background:

• Structural variations (SVs) significantly influence genome evolution, disease, and phenotypic diversity.
• Existing SV callers, often optimized for human genomes, require validation for avian species due to genomic differences.
• Rigorous evaluation of SV callers using real data and validated SVs is essential for accurate genomic analysis.

Purpose of the Study:

• To comprehensively assess the performance of ten widely used SV callers on avian population genomic data.
• To evaluate SV caller accuracy across different SV types and sizes, including insertions, deletions, and copy number variations.
• To provide practical guidance for SV detection in avian research.

Main Methods:

• Performance evaluation of ten SV callers using population-level, real genomic data.
• Utilized validated common types of SVs for benchmarking.
• Analyzed SV detection accuracy based on SV type, size, and required read depth.

Main Results:

• SV caller performance varied significantly across different SV types and sizes.
• GRIDSS, Lumpy, Wham, and Manta demonstrated superior detection accuracy; Pindel excelled at small SVs; CNVnator and CNVkit identified medium to large copy number variations.
• Combination calling strategies were not recommended due to poor consistency; high read depth (≥50×) is necessary for >80% SV detection; insertion detection, especially >150 bp, was challenging for all tools.

Conclusions:

• Emphasizes the critical need for evaluating SV callers with real sequencing data and validated SVs, not solely simulated data.
• Highlights the differential performance of SV callers, guiding tool selection for specific avian genomic studies.
• Offers practical recommendations for optimizing SV detection strategies in avian research.