Inferring Heterozygosity from Ancient and Low Coverage Genomes

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• Accurate genetic diversity quantification is crucial but challenging with low-coverage or degraded DNA, common in ancient samples.
• Existing methods often rely on high coverage or strict assumptions, limiting their applicability to non-model organisms and ancient genomes.

Purpose of the Study:

• To develop a probabilistic method for accurate heterozygosity inference from low-coverage sequencing data.
• To create a robust recalibration method for sequencing error rates, particularly in the presence of postmortem damage.
• To apply these methods to analyze genetic diversity in ancient human populations.

Main Methods:

• A probabilistic method to infer heterozygosity from low-coverage sequencing data, relaxing the infinite sites assumption and accommodating sequencing errors and postmortem damage.
• A recalibration method for sequencing error rates using haploid data, integrating over unknown genotypes and effective even with low coverage.
• Application of the developed methods to infer genome-wide diversity patterns in ancient human samples.

Main Results:

• The new method accurately estimates heterozygosity from low-coverage data (average coverage of 1x) down to 10^-5 within 1 Mbp windows.
• Accurate recalibration of sequencing error rates is achievable with a few megabasepairs of haploid data, even at low coverages (0.1x).
• Analysis of ancient humans revealed diversity patterns comparable to modern humans in 3000-5000-year-old samples, while European hunter-gatherers showed lower diversity and distinct genomic patterns.

Conclusions:

• The developed probabilistic method offers accurate genetic diversity estimation from low-coverage and ancient DNA.
• The recalibration method improves accuracy by accounting for sequencing errors and postmortem damage.
• The study provides insights into the genetic diversity and population structure of ancient European populations.