Best practices for analyzing imputed genotypes from low-pass sequencing in dogs

Area of Science:

• Genomics
• Canine genetics
• Bioinformatics

Background:

• DNA array-based genome-wide association studies (GWAS) provide low-cost genotypes but lack resolution.
• SNP chip technology is limited by pre-selected markers.
• Low-pass whole genome sequencing (WGS) with imputation is a promising alternative for comprehensive genetic analysis.

Purpose of the Study:

• To assess the accuracy and performance of low-pass WGS followed by imputation in dogs.
• To optimize variant quality filtering for imputed genotypes.
• To provide guidelines for low-pass WGS study design and data processing in canine genetics.

Main Methods:

• Downsampled 97 high-coverage WGS datasets (> 15×) to ~1× coverage to simulate low-pass WGS.
• Utilized a reference panel of 676 dogs from 91 breeds for imputation.
• Compared imputed genotypes to a high-coverage WGS truth set and optimized filtering strategies.

Main Results:

• Optimized filtering retained ~80% of 14 million imputed sites, reducing error rate from 3.0% to 1.5%.
• Seven million sites remained with minor allele frequency >5% and average imputation quality score of 0.95.
• Simulated GWAS indicated imputation errors minimally impacted medium-to-large effect sizes but affected small effect sizes.

Conclusions:

• Low-pass WGS with imputation is a viable and accurate method for canine genetic studies.
• Optimized filtering strategies are crucial for maximizing data quality and utility.
• This approach provides a cost-effective pathway for large-scale canine genetic research and GWAS.