GGTyper: genotyping complex structural variants using short-read sequencing data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Complex structural variants (SVs) are significant contributors to human diversity and Mendelian diseases.
• Current analytical tools for complex SVs are limited, hindering comprehensive genomic analysis.

Purpose of the Study:

• To develop and validate a novel computational approach for accurate genotyping of complex structural variants (SVs) in short-read sequenced genomes.
• To create a method that predicts genotyping difficulty, enabling efficient data requirements assessment for reliable variant calls.

Main Methods:

• Developed a computational approach to calculate genotype-specific probability distributions for aligned read pair properties.
• Utilized these distributions to determine the most probable genotype for observed read pairs.
• Integrated a genotyping difficulty prediction metric based on these probability distributions.

Main Results:

• The new approach demonstrated superior performance over existing genotypers on both simulated and real genomic data.
• Achieved high concordance with population genetics principles and inheritance patterns across 7829 human genomes.
• Demonstrated a strong correlation between predicted genotyping difficulty and actual precision, with low computational resource requirements.

Conclusions:

• The developed method provides a robust and efficient solution for complex SV genotyping in large-scale genomic studies.
• This approach is well-suited for diverse biomedical applications, from small cohorts to very large population-scale sequencing projects.
• The availability of source code facilitates broader adoption and further development in the field.