Deep-time paleogenomics and the limits of DNA survival

Abstract

Although most ancient DNA studies have focused on the last 50,000 years, paleogenomic approaches can now reach into the early Pleistocene, an epoch of repeated environmental changes that shaped present-day biodiversity. Emerging deep-time genomic transects, including from DNA preserved in sediments, will enable inference of adaptive evolution, discovery of unrecognized species, and exploration of how glaciations, volcanism, and paleomagnetic reversals shaped demography and community composition. In this Review, we explore the state-of-the-art in paleogenomics and discuss key challenges, including technical limitations, evolutionary divergence and associated biases, and the need for more precise dating of remains and sediments. We conclude that with improvements in laboratory and computational methods, the emerging field of deep-time paleogenomics will expand the range of questions addressable using ancient DNA.

Fig. 1. The temporal distribution of ancient DNA studies to date highlights gaps and opportunities for deep-time paleogenomics and sedimentary ancient DNA.

Circles in orange are non-human animal paleogenomes, in blue are hominin paleogenomes, and in brown are sedimentary ancient DNA records. Most ancient DNA studies fall within the last 50 ka and the most recent glacial cycle. The climate curve is based on benthic δ¹⁸-Oxygen measurements (per mil, %_o, LR04 stack from (42). Sedimentary ancient DNA data are from the AncientMetagenomeDir (v23.06.0, 58) and von Eggers et al. (v1, https://doi.org/10.5281/zenodo.6847522), with metabarcoding records older than one million years excluded. Paleogenomes older than 100 ka are annotated with a silhouette of the study taxon, with the deep-time paleogenomes including a 130 ka steppe bison (36); 330 ka collared lemming (40); 360 ka cave bear (9); 430 ka cave bear and hominin (35, 59); 700 ka horse (8); and 700 ka, 1.1 Ma, and 1.2 Ma mammoths (10). Silhouettes are from PhyloPic (https://beta.phylopic.org/) and are in the public domain with credits to Zimices (mammoth, two bison) and Robert Bruce Horsfall (horse). LP: Late Pleistocene; IG: Interglacial; G: Glacial.

Fig. 2. DNA fragmentation and degradation begins after death and continues until fragments are too short to be useful.

(A) The integrity of megabase length DNA is maintained by a cell's enzymatic repair machinery and, in eukaryotic genomes, packaged in histone-DNA complexes. (B) Following death, repair stops and DNA damage begins to accumulate. Nucleases and microorganisms cleave DNA in labile regions between nucleosomes and when the DNA backbone faces away from histones. (C) Over time, chemical damage also accumulates. Cytosine bases are converted to uracil and methylated cytosines are converted to thymines (by deamination). Cytosines are particularly vulnerable to deamination in single-stranded regions such as in overhanging regions at DNA termini, but deamination is possible in some double-stranded contexts. Fragmentation occurs after the loss of purine bases (depurination), creating abasic sites that can be cleaved by β elimination. Depurination and β elimination create a region of single-stranded DNA, which leaves cytosines vulnerable to deamination.. (D) Given enough time, DNA molecules will become too short to be identifiable. (E) A summary of base and mismatch frequencies along the initial 15 5’ and 3’ bases of reads generated using a single-stranded DNA library protocol. Depurination leads to overrepresentation of adenine and guanine bases adjacent to strand breaks. C-to-T mismatches are elevated near read ends and observed throughout damaged reads. While 3’ G-to-A mismatches are observed in double-stranded libraries, single-stranded libraries show a C-to-T signal at both ends by retaining the native termini of the molecules.

Fig. 3. Deep-time paleogenomes provided new understanding of the evolutionary history of mammoths.

Paleontological hypotheses assumed that the M. columbi lineage evolved after early divergence from M. primigenius (A), however, isolation of a deep-time paleogenome from the Krestova mammoth (blue circle) revealed that M. columbi emerged more recently and following admixture with the Krestova mammoth lineage (B).