Detection of mosaic and population-level structural variants with Sniffles2

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Accurate identification of structural variations (SVs) is crucial for understanding complex genomic alterations.
• Long-read sequencing technologies offer superior accuracy for detecting these variations compared to short reads.
• Existing SV callers face challenges in speed, accuracy, and handling complex genomic structures.

Purpose of the Study:

• To present Sniffles2, an improved long-read SV caller.
• To enhance the speed and accuracy of structural variation detection.
• To enable comprehensive SV analysis from family-level to population-level studies and detect mosaic SVs.

Main Methods:

• Implementation of repeat-aware clustering in Sniffles2.
• Utilizing a fast consensus sequence generation.
• Incorporating coverage-adaptive filtering for SV detection.
• Application across various sequencing coverages (5-50×) and technologies (ONT, HiFi).

Main Results:

• Sniffles2 demonstrates 11.8x speed improvement and 29% higher accuracy over state-of-the-art SV callers.
• Successfully identified causative SVs, including complex alleles, in probands near the MECP2 gene.
• Enabled detection of mosaic SVs in bulk long-read brain tissue data from a multiple system atrophy patient.
• Revealed diverse mosaic SVs in the cingulate cortex impacting neuronal function genes and repetitive elements.

Conclusions:

• Sniffles2 represents a significant advancement in structural variation calling using long-read sequencing.
• The tool provides accurate, fast, and comprehensive SV detection, including complex and mosaic alterations.
• Sniffles2 facilitates deeper insights into genetic diseases and neurological disorders through detailed SV analysis.