CONGA: Copy number variation genotyping in ancient genomes and low-coverage sequencing data

To date, ancient genome analyses have been largely confined to the study of single nucleotide polymorphisms (SNPs). Copy number variants (CNVs) are a major contributor of disease and of evolutionary adaptation, but identifying CNVs in ancient shotgun-sequenced genomes is hampered by (i) most published genomes being <1x coverage, (ii) ancient DNA fragments being typically <80 bps. These characteristics preclude state-of-the-art CNV detection software to be effectively applied to ancient genomes. Here we present CONGA, an algorithm tailored for genotyping deletion and duplication events in genomes with low depths of coverage. Simulations and down-sampling experiments show that CONGA can genotype deletions >1 kbps with F-scores >0.75 at >=1x, and distinguish between heterozygous and homozygous states. Using CONGA, we analyse deletion events at 10,018 loci in 56 ancient human genomes spanning the last 50,000 years, with coverages 0.4x-26x. We show that inter-individual genetic diversity measured using deletions and SNPs are highly correlated, as in modern-day genomes, confirming that deletion frequencies broadly reflect demographic history. We also identify signatures of strong purifying selection on deletions in ancient-genomes, such as an excess of singletons compared to those in SNPs. CONGA paves the way for systematic studies of drift, mutation load, and adaptation in ancient and modern-day gene pools through the lens of CNVs.