Imputation of ancient human genomes.

Area of Science:

• Genomics
• Ancient DNA Analysis
• Bioinformatics

Background:

• Ancient genomes often suffer from low coverage due to DNA degradation and microbial contamination.
• Low-coverage genomes present challenges for accurate genotype calling and downstream analyses.
• Genotype imputation is a potential method to enhance data quality for ancient DNA.

Purpose of the Study:

• To evaluate the accuracy and potential biases of genotype imputation in ancient DNA.
• To determine the minimum coverage required for reliable imputation in ancient genomes.
• To assess the impact of imputation on downstream population genetic analyses.

Main Methods:

• Re-sequencing of an ancient trio (mother, father, son).
• Downsampling and imputation of 43 ancient genomes (42 high-coverage).
• Assessment of imputation accuracy across ancestries, time, coverage depth, and sequencing technology.
• Validation using trio data and Mendelian inheritance rules.
• Comparison of downstream analyses (PCA, clustering, ROH) between imputed and high-coverage genomes.

Main Results:

• Ancient DNA imputation accuracy is comparable to modern DNA imputation.
• At 1x coverage, 36 out of 42 genomes showed low imputation error rates (<5%), with higher errors in African genomes.
• Downstream analyses yielded similar results for imputed and high-coverage genomes from 0.5x coverage, with exceptions for African genomes.
• Imputation is reliable for most populations at coverage depths as low as 0.5x.

Conclusions:

• Genotype imputation is a robust and reliable method for improving the analysis of low-coverage ancient DNA.
• Imputation significantly enhances the utility of ancient genomes, enabling more accurate genotype calling and downstream analyses.
• The findings support the broader application of imputation in ancient DNA research, particularly for populations with limited sample availability or degraded DNA.