Mining ancient microbiomes using selective enrichment of damaged DNA molecules

Abstract

Background: The identification of bona fide microbial taxa in microbiomes derived from ancient and historical samples is complicated by the unavoidable mixture between DNA from ante- and post-mortem microbial colonizers. One possibility to distinguish between these sources of microbial DNA is querying for the presence of age-associated degradation patterns typical of ancient DNA (aDNA). The presence of uracils, resulting from cytosine deamination, has been detected ubiquitously in aDNA retrieved from diverse sources, and used as an authentication criterion. Here, we employ a library preparation method that separates molecules that carry uracils from those that do not for a set of samples that includes Neandertal remains, herbarium specimens and archaeological plant remains.

Results: We show that sequencing DNA libraries enriched in molecules carrying uracils effectively amplifies age associated degradation patterns in microbial mixtures of ancient and historical origin. This facilitates the discovery of authentic ancient microbial taxa in cases where degradation patterns are difficult to detect due to large sequence divergence in microbial mixtures. Additionally, the relative enrichment of taxa in the uracil enriched fraction can help to identify bona fide ancient microbial taxa that could be missed using a more targeted approach.

Conclusions: Our experiments show, that in addition to its use in enriching authentic endogenous DNA of organisms of interest, the selective enrichment of damaged DNA molecules can be a valuable tool in the discovery of ancient microbial taxa.

Keywords: Ancient DNA; Authentication; Metagenomics.

Fig. 1

Relative enrichment, taxonomic assignment and substitution profiles of Neandertal-derived U-selected libraries. a Relative enrichment (number of reads) in the U-enriched relative to the U-depleted fraction from Vindija Neandertal assigned to the phyla Actinobacteria and Proteobacteria, as well as to Homo sapiens. b Relative enrichment (number of reads) in the U-enriched relative to the U-depleted fraction from Sidrón Neandertal assigned to the phyla Actinobacteria and Proteobacteria, as well as to Homo sapiens. c Taxonomic tree of reads from Sidrón Neandertal assigned to different taxonomic levels. The size of the circle represents the amount of reads assigned to the node displayed in the tree or to any taxonomic level below it. Assignments to the phyla Actinobacteria and Proteobacteria, as well as the species Streptosporangium roseum and Homo sapiens are named in the taxonomic tree. Other taxonomic groups are either unlabeled or removed from the tree to increase clarity. d Cytosine to Thymine substitutions at the 5′ end of reads aligned to S. roseum from the Sidrón Neandertal U-selected library (U-enriched and U-depleted fractions)

Fig. 2

Patterns of cytosine to thymine (C-to-T) substitutions at the 5′ end of plant- and Pseudomonas-derived reads. a C-to-T substitutions at the 5′ end of Solanum tuberosum sample KM177500 for a non-selected and U-selected library (U-enriched and U-depleted fractions). b Distributions of C-to-T substitution percentage at first base (5′ end) for non-selected and U-selected libraries (U-enriched and U-depleted fractions). Median values are denoted as black lines and points show the original value for each individual sample. c Substitution patterns at the 5′ end of Pseudomonas syringae and Pseudomonas rhizosphaerae mapped reads from a non-selected library from a Solanum tuberosum sample KM177500. d Cytosine to Thymine substitutions at the 5′ end of P. syringae and P. rhizosphaerae mapped reads from a U-selected library (U-enriched and U-depleted fractions) from a Solanum tuberosum sample KM177500

Fig. 3

Characterization of the bacterium Pantoea vagans identified in Zea mays and Solanum tuberosum samples. a Taxonomic tree of reads from Solanum tuberosum and Zea mays assigned to different taxonomic levels. The size of the circle represents the amount of reads assigned to the node displayed in the tree or to any taxonomic level below it. S. tuberosum- and Z. mays-derived reads are shown in green and orange, respectively. b Cytosine to Thymine substitutions at the 5′ end of P. vagans for U-selected libraries (U-enriched and U-depleted fractions) from Z. mays and S. tuberosum. c Principal component analysis of P. vagans from Z. mays and S. tuberosum samples, as well as nine publicly available genomes, based on single nucleotide polymorphisms. Numbers in axis labels indicate the percentage of the variance explained by each principal component (PC)