GraphTyper2 enables population-scale genotyping of structural variation using pangenome graphs

Area of Science:

• Genomics
• Bioinformatics
• Human Genetics

Background:

• Analysis of human genome sequence diversity is crucial for genetic studies.
• Structural variants (SVs) are often overlooked in sequence analyses despite their significant genomic impact.

Purpose of the Study:

• To introduce GraphTyper2, a novel tool for genotyping structural variants (SVs) and small variants using pangenome graphs and short-read sequencing data.
• To enhance the sensitivity and accuracy of SV detection in large-scale human genome analyses.

Main Methods:

• Utilized pangenome graphs for variant genotyping of both SVs and small variants.
• Incorporated public assembly data into the analysis pipeline to improve sensitivity.
• Validated SV genotyping accuracy using family segregation data and long-read sequencing.

Main Results:

• GraphTyper2 demonstrates sensitive SV genotyping, validated by benchmark datasets and family segregation.
• Incorporating public assembly data significantly boosted sensitivity, particularly for large insertions.
• Successfully genotyped 60 million small variants and 0.5 million SVs across 49,962 Icelandic genomes, including 80,000 high-confidence SVs.

Conclusions:

• GraphTyper2 provides an accurate and sensitive method for simultaneously genotyping small variants and SVs in large cohorts.
• The tool's ability to leverage pangenome graphs and public data advances the comprehensive analysis of human genome diversity.