Genotype imputation from low-coverage data for medical and population genetic analyses.

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• Low-pass sequencing data offers genomic analysis potential but faces challenges.
• Conventional quality filters may be inadequate for ultra-low coverage pooled sequences.

Purpose of the Study:

• To assess the impact of quality filters on genetic ancestry and Polygenic Score (PGS) estimation from low-pass sequencing data.
• To introduce and evaluate a novel filtering approach (GDI) for enhancing genotype quality.

Main Methods:

• Imputed 32,769 low-pass genome-wide sequences (LPS) from NIPS data.
• Applied QUILT and GLIMPSE2 imputation tools.
• Developed and tested the GDI filter combining genotype probability, alternate allele dosage, and INFO score.

Main Results:

• Imputation tools achieved high accuracy for broad ancestry mapping but not high-resolution PCA without filters.
• The GDI filter improved genotype quality to be adequate for PCA.
• GDI filtering did not enhance PGS prediction for height, unlike its effect on PCA.

Conclusions:

• Imputed LPS data is valuable for biomedical and population genetics.
• Downstream analysis tools require individual consideration of imputation quality thresholds and filtering needs.