The Allen Ancient DNA Resource (AADR) a curated compendium of ancient human genomes.
Mallick S, Micco A, Mah M, Ringbauer H, Lazaridis I, Olalde I, Patterson N, Reich D. · 2024
Read The Allen Ancient DNA Resource (AADR) a curated compendium of ancient human genomes. by Mallick S, Micco A, Mah M, Ringbauer H, Lazaridis I, Olalde I, Patterson N, Reich D. (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc10858950.txt.
Introduction
Mallick and co-authors describe a shared archive of ancient human genomes in this 2024 paper. Ancient DNA is genetic material read from the remains of people who died long ago. More than 200 papers have published genome-wide data from such remains. Raw data are mostly public, but formats differ from study to study, which makes joint work hard.
The Allen Ancient DNA Resource, kept since 2019, cleans and unifies those records. It holds data at over a million DNA spots for each person. By late 2022 it passed 10,000 ancient people, and it had seen six public releases. The note serves as a citable record of the archive. The authors aim to give the field one versioned source that all can download and trust.
Catalog record
| CREATOR | Mallick S, Micco A, Mah M, Ringbauer H, Lazaridis I, Olalde I, Patterson N, Reich D. |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Allen Ancient DNA Resource (AADR) a curated compendium of ancient human genomes By Swapan Mallick, Adam Micco, Matthew Mah, Harald Ringbauer, Iosif Lazaridis, Iñigo Olalde, Nick Patterson, David Reich Published in: Scientific Data DOI: 10.1038/s41597-024-03031-7 ABSTRACT Abstract More than two hundred papers have reported genome-wide data from ancient humans. While the raw data for the vast majority are fully publicly available testifying to the commitment of the paleogenomics community to open data, formats for both raw data and meta-data differ. There is thus a need for uniform curation and a centralized, version-controlled compendium that researchers can download, analyze, and reference. Since 2019, we have been maintaining the Allen Ancient DNA Resource (AADR), which aims to provide an up-to-date, curated version of the world’s published ancient human DNA data, represented at more than a million single nucleotide polymorphisms (SNPs) at which almost all ancient individuals have been assayed. The AADR has gone through six public releases at the time of writing and review of this manuscript, and crossed the threshold of >10,000 individuals with published genome-wide ancient DNA data at the end of 2022. This note is intended as a citable descriptor of the AADR. Subject terms: Genetic variation, Genetic variation BACKGROUND & SUMMARY The first genome-wide ancient DNA data were published in 20101–3. However, it was only in 2015 with the advent of large-scale studies of Holocene genomes, in-solution enrichment of ancient DNA libraries for targeted single nucleotide polymorphisms (SNPs)4–6, and the introduction of automated protocols and liquid handling robots for processing of ancient DNA libraries7,8, that the number of individuals with genome-wide data began to increase rapidly. Between 2010 and 2014, data from an average of about 10 individuals with genome-wide data were published each year. Between 2015 and 2017, the numbers increased to about 200 annually. Since 2018, data from thousands of individuals have been published every year (Fig. 1). About 67% of the data are from Europe and Russia, a proportion that has held relatively steady since the beginning of the field of ancient DNA. The proportion of data from East Asia has grown from about 1% of all data in 2015 to 8% today. The proportion of data from Africa has grown from none in 2014 to 3% today (Fig. 1). [Figure: Growth in world’s published human genome-wide ancient DNA data. (A) By year of publication (broken down by geography). (B) By date (color and symbol both indicate geographic location). (C) By geography (using same color and symbol scheme as in previous panel).] A challenge in analyzing ancient DNA data is that it has been reported over hundreds of independent studies. Thus, while raw sequence data for more than 99% of individuals9 are fully available in public repositories such as ENA10 and SRA11, the uploaded data exist in diverse formats, as do the meta-data such as archaeological, chronological, and geographic information. Some resources exist which consolidate subsets of publicly available ancient DNA data, including a Y-chromosome database with assembled information from nearly two thousand ancient Eurasian individuals12, a mitochondrial DNA database with more than two thousand individuals13, and the Online Ancient Genome Repository14 which copies publicly available data and encapsulates each dataset into an archived tar file. However, none of these provide a regularly curated dataset that attempts to include all published data in an easily co-analyzable format, such as a single genotype file with complete annotations in a single tab-delimited form. Sources of data To bring data generated outside our own laboratory into the AADR, we usually start with available sequences from a public repository, most often the European Nucleotide Archive (https://www.ebi.ac.uk/ena), following accession numbers given in the published papers. In some cases we start with alternatively formatted versions that we request directly from the authors. All source articles are cited in the reference list of this paper. For data generated in our laboratory, we start from our own raw sequence files, which are the basis for data uploaded to established public repositories. METHODS The raw data generated outside our laboratory come in diverse formats, usually fastq files (for raw sequence data) or bams (for either unaligned reads or reads aligned to a reference genome)15. A challenge is that there can be considerable variation in fastq and bam files, reflecting the formatting, filtering and processing choices made by researchers in generating data. This includes: - Base calls and associated quality scores in raw sequences are often modified by the researchers who generated the data. One common modification is to recalibrate base quality scores16. Another modification is to ignore information from the ends of sequences, either by masking terminal bases in the sequences that are uploaded and marking them as “N”, or clipping (removing) them altogether17. This reduces error rates associated with cytosine deamination typical of ancient DNA data. However, it also means that users cannot make choices about whether to use the valuable data that have been masked and clipped (such as sites unaffected by deamination). In addition, this procedure has the effect of making it difficult to identify damaged molecules which are a strong indicator that those molecules indeed are ancient and not derived from some potential contaminating modern human source. - Sequences may be aligned to different human reference genomes, typically hg19, hs37d5, or hg20, each with their own unique coordinate systems. To build a homogeneous dataset, we therefore have to map to a unified coordinate system, currently based on hg1918–20. A further challenge is that chromosomes may have inconsistent naming conventions (for example ‘chr1’ v. ‘1’, or ‘chrMT’ v. ‘MT’ v. ‘chrM’), or the sorting order of chromosomes can differ. This results in practical difficulties in merging datasets. - Data may be deposited either (i) by library, or (ii) by-individual with multiple libraries in a single file. If data are deposited by library, then it may be necessary to identify and perform a merging step. There are pitfalls that arise in such merging, as in some cases “readgroup” names (a tag which groups reads together) are the same across individuals, and so joint processing of many individuals can inadvertently lead to in-silico contamination. Filtering of data To add data to the AADR, we manually process the dataset from each individual, tailoring the processing procedure according to the characteristics of the data. We create a bam file aligned to the hg19 genome reference sequence. The bam files used to generate AADR constitute tens of terabytes in size altogether. We process these bams to produce genotypes at a set of about 1.23 million SNPs that have been assayed for nearly all published individuals with ancient DNA data. For the great majority of ancient DNA datasets, the genotypes are “pseudohaploid”, meaning that we represent the individual by a randomly sampled sequence at each analyzed position. For the small fraction of individuals for whom coverage is sufficient to allow full genotyping, we also release diploid genotypes21,22. Combining datasets To increase the usefulness of the AADR, we have added into the AADR data from diverse modern humans, including shotgun sequencing data from sets of individuals included within the 1000 Genomes Project23, the Simons Genome Diversity Project24, and the Human Genome Diversity Project25 To integrate these data, we had to address challenges of different reference genomes (for