The Allen Ancient DNA Resource (AADR): 1240K, Human Origins and sample labels

The Allen Ancient DNA Resource (AADR) is a curated dataset of published ancient and present-day human genomes, released in versions by David Reich's laboratory at Harvard. It puts genomes from hundreds of studies on one set of SNPs, with one annotation table, so they can be analysed together. Most qpAdm, f-statistic and PCA results in the ancient DNA literature use it.

Mallick et al. 2024 describe the resource in Scientific Data. Each release adds the genomes published since the one before.

Two releases: 1240K and Human Origins

Each AADR version comes in two main releases, by SNP set:

• 1240K: about 1.2 million SNPs, the set that the in-solution capture reagent of the same name targets. Most ancient genomes are captured or shotgun-sequenced and genotyped at these SNPs. Use 1240K for ancient-only analyses and for f-statistics, where more SNPs means smaller standard errors.
• Human Origins (HO): about 600,000 SNPs from the Affymetrix Human Origins array (Patterson et al. 2012), which was used to genotype many present-day populations worldwide. The HO release has the ancient genomes at these SNPs plus far more present-day populations. Use HO for PCA and for models that need present-day reference groups.

Nexju uses AADR v66 with three published supplements added: 23,870 genomes in its 1240K dataset and 28,375 in its Human Origins dataset. Nexju also uses the AADR v66 2M-compatibility patch for merges with consumer DNA files that overlap it well.

The file format

The AADR ships in EIGENSTRAT (packed) format:

• .geno: the genotypes, one row per SNP and one column per genome;
• .snp: the SNPs, with chromosome, genetic and physical positions and alleles;
• .ind: the genomes, with sex and population label;
• .anno: the annotation table, one row per genome.

ADMIXTOOLS and EIGENSOFT read this format directly. PLINK and ADMIXTOOLS 2 can convert it.

The annotation table

The .anno file is the most useful part for choosing populations. For each genome it lists, among other columns:

• Genetic ID and Master ID: the genome's identifier, and the person's identifier across libraries and releases.
• Group ID: the population label used in analyses.
• Date: mean date in years before present (BP, before 1950), and how it was obtained (radiocarbon or archaeological context).
• Locality, country, latitude and longitude.
• Data source and coverage: how the genome was sequenced and how many 1240K SNPs it covers.
• Assessment: a quality flag, such as PASS, or QUESTIONABLE with a reason.

Reading group labels

Group labels follow a pattern: country, site, period and culture, joined by underscores. For example:

• Russia_Samara_EBA_Yamnaya: Yamnaya culture, Early Bronze Age, from Samara in Russia.
• Luxembourg_Loschbour_Mesolithic: the Mesolithic hunter-gatherer from Loschbour.
• Turkey_N: Neolithic Anatolians.
• Mbuti, Sardinian, English: present-day populations, named by group.

Suffixes mark special cases:

• -o: an outlier, a genome whose ancestry differs from the rest of its site. A suffix after it, such as -oWHG, says towards which ancestry.
• -1, -2: sub-groups of a site with different ancestry.

Genetic IDs carry their own suffixes, such as .SG for shotgun-sequenced data and .DG for diploid genotypes from high-coverage genomes. One person can appear more than once, from different libraries or SNP sets; check the Master ID before you count genomes.

Pseudohaploid data

Most ancient genomes have low coverage, so a heterozygous site cannot be called reliably. The AADR reports them as pseudohaploid: one read is sampled at each SNP, and the genome is recorded as homozygous for that allele. f-statistics and smartpca's projection handle pseudohaploid data, but count each such genome as carrying half the information of a diploid one.

Using the AADR

The AADR is free to download from the Harvard Dataverse. Nexju has it loaded: you search populations and samples by name, period, region and coverage, save the ones you choose as a dataset, and run qpAdm, f-statistics and PCA on them in the browser.

References

• Mallick S. et al. 2024. The Allen Ancient DNA Resource (AADR) a curated compendium of ancient human genomes. Scientific Data 11. doi:10.1038/s41597-024-03031-7
• Patterson N. et al. 2012. Ancient admixture in human history. Genetics 192. doi:10.1534/genetics.112.145037
• Mathieson I. et al. 2015. Genome-wide patterns of selection in 230 ancient Eurasians. Nature 528. doi:10.1038/nature16152