Limits...
Early divergent strains of Yersinia pestis in Eurasia 5,000 years ago.

Rasmussen S, Allentoft ME, Nielsen K, Orlando L, Sikora M, Sjögren KG, Pedersen AG, Schubert M, Van Dam A, Kapel CM, Nielsen HB, Brunak S, Avetisyan P, Epimakhov A, Khalyapin MV, Gnuni A, Kriiska A, Lasak I, Metspalu M, Moiseyev V, Gromov A, Pokutta D, Saag L, Varul L, Yepiskoposyan L, Sicheritz-Pontén T, Foley RA, Lahr MM, Nielsen R, Kristiansen K, Willerslev E - Cell (2015)

Bottom Line: How and when it originated remains contentious.We also identify a temporal sequence of genetic changes that lead to increased virulence and the emergence of the bubonic plague.Our results show that plague infection was endemic in the human populations of Eurasia at least 3,000 years before any historical recordings of pandemics.

View Article: PubMed Central - PubMed

Affiliation: Center for Biological Sequence Analysis, Department of Systems Biology, Technical University of Denmark, Kemitorvet, Building 208, 2800 Kongens Lyngby, Denmark.

Show MeSH

Related in: MedlinePlus

Authenticity of Y. pestis DNA(A) DNA damage patterns for RISE505 and RISE509. The frequencies of all possible mismatches observed between the Y. pestis CO92 chromosome and the reads are reported in gray as a function of distance from 5′ (left panel, first 25 nucleotides sequenced) and distance to 3′ (right panel, last 25 nucleotides). The typical DNA damage mutations C>T (5′) and G>A (3′) are reported in red and blue, respectively.(B) Ancient DNA damage patterns (n = 7) of the reads aligned to the CO92 chromosome and the Y. pestis associated plasmids pMT1, pCD1 and pPCP1. The boxplots show the distribution of C-T damage in the 5′ of the reads. The lower and upper hinges of the boxes correspond to the 25th and 75th percentiles, the whiskers represent the 1.5 inter-quartile range (IQR) extending from the hinges, and the dots represent outliers from these.(C) DNA fragment length distributions from RISE505 and RISE509 samples representing both the Y. pestis DNA and the DNA of the human host. The declining part of the distributions is fitted to an exponential model (red).(D) Linear correlation (red) between the decay constant in the DNA of the human host and the associated Y. pestis DNA extracted from the same individual (R2 = 0.55, p = 0.055). The decay constant (λ) describes the damage fraction (i.e., the fraction of broken bonds on the DNA strand).(E) Distribution of edit distance of high quality reads from RISE505 and RISE509 samples mapped to either Y. pestis (dark gray) or Y. pseudotuberculosis (light gray) reference genomes. The reads have a higher affinity to Y. pestis than to Y. pseudotuberculosis.(F) Plots of actual coverage versus expected coverage for the 101 screened samples. Expected coverage was computed taking into account read length distributions, mappable fractions of reference sequences, and the deletions in pMT1 for some of the samples. Samples assumed to contain Y. pestis are shown in blue and RISE392 that is classified as not Y. pestis appears is shown in red. See also Figure S1 and S2, Table S3.
© Copyright Policy - CC BY
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4644222&req=5

fig3: Authenticity of Y. pestis DNA(A) DNA damage patterns for RISE505 and RISE509. The frequencies of all possible mismatches observed between the Y. pestis CO92 chromosome and the reads are reported in gray as a function of distance from 5′ (left panel, first 25 nucleotides sequenced) and distance to 3′ (right panel, last 25 nucleotides). The typical DNA damage mutations C>T (5′) and G>A (3′) are reported in red and blue, respectively.(B) Ancient DNA damage patterns (n = 7) of the reads aligned to the CO92 chromosome and the Y. pestis associated plasmids pMT1, pCD1 and pPCP1. The boxplots show the distribution of C-T damage in the 5′ of the reads. The lower and upper hinges of the boxes correspond to the 25th and 75th percentiles, the whiskers represent the 1.5 inter-quartile range (IQR) extending from the hinges, and the dots represent outliers from these.(C) DNA fragment length distributions from RISE505 and RISE509 samples representing both the Y. pestis DNA and the DNA of the human host. The declining part of the distributions is fitted to an exponential model (red).(D) Linear correlation (red) between the decay constant in the DNA of the human host and the associated Y. pestis DNA extracted from the same individual (R2 = 0.55, p = 0.055). The decay constant (λ) describes the damage fraction (i.e., the fraction of broken bonds on the DNA strand).(E) Distribution of edit distance of high quality reads from RISE505 and RISE509 samples mapped to either Y. pestis (dark gray) or Y. pseudotuberculosis (light gray) reference genomes. The reads have a higher affinity to Y. pestis than to Y. pseudotuberculosis.(F) Plots of actual coverage versus expected coverage for the 101 screened samples. Expected coverage was computed taking into account read length distributions, mappable fractions of reference sequences, and the deletions in pMT1 for some of the samples. Samples assumed to contain Y. pestis are shown in blue and RISE392 that is classified as not Y. pestis appears is shown in red. See also Figure S1 and S2, Table S3.

Mentions: Besides applying standard precautions for working with ancient DNA (Willerslev and Cooper, 2005), the authenticity of our findings are supported by the following observations: (1) The Y. pestis sequences were identified in significant amounts in shotgun data from eight of 101 samples, showing that this finding is not due to a ubiquitous contaminant in our lab or in the reagents. Indeed, further analysis showed that one of these eight was most likely not Y. pestis. We also sequenced all negative DNA extraction controls and found no signs of Y. pestis DNA in these (Table S3). (2) Consistent with an ancient origin, the Y. pestis reads were highly fragmented, with average read lengths of 43–65 bp (Table S3) and also displayed clear signs of C-T deamination damage at the 5′ termini typical of ancient DNA (Figure 3, Figure S1). Because the plasmids are central for discriminating between Y. pestis and Y. pseudotuberculosis, we tested separately for DNA damage patterns for the chromosome and for each of the plasmids. For the seven samples, we observe similar patterns of DNA damage for chromosome and plasmid sequences (Figure 3, Figure S1). (3) We observe correlated DNA degradation patterns when comparing DNA degradation in the Y. pestis sequences and the human sequences from the host individual. Given that DNA decay can be described as a rate process (Allentoft et al., 2012), this suggests that the DNA molecules of the pathogen and the human host have a similar age (Figure 3, Figure S1, Table S3 and Supplemental Experimental Procedures). (4) Because of the high sequence similarity between Y. pestis and Y. pseudotuberculosis, we mapped all reads both to the Y. pestis CO92 and to the Y. pseudotuberculosis IP32953 reference genomes (Chain et al., 2004). Consistent with being Y. pestis, the seven investigated samples displayed more reads matching perfectly (edit distance = 0) toward Y. pestis (Figure 3, Figure S2). One sample (RISE392) was most likely not Y. pestis based on this criterion. (5) A naive Bayesian classifier trained on known genomes predicts the seven samples to be Y. pestis with 100% posterior probability, while RISE392 is predicted to have 0% probability of being Y. pestis (Figure S2, Table S3). (6) If the DNA was from other organisms than Y. pestis, we would expect the reads to be more frequently associated with either highly conserved or low-complexity regions. However, we find the reads to be distributed across the entire genome (Figure S2), and comparison of actual coverage versus the coverage that would be expected from read length distributions and mappability of the reference sequences are also in agreement for the seven samples (Figure 3). (7) In a maximum likelihood phylogeny, the recovered Y. pestis genomic sequences of RISE505 and RISE509 are clearly within the Y. pestis clade and basal to all contemporary Y. pestis strains (Figure 4) (see below).


Early divergent strains of Yersinia pestis in Eurasia 5,000 years ago.

Rasmussen S, Allentoft ME, Nielsen K, Orlando L, Sikora M, Sjögren KG, Pedersen AG, Schubert M, Van Dam A, Kapel CM, Nielsen HB, Brunak S, Avetisyan P, Epimakhov A, Khalyapin MV, Gnuni A, Kriiska A, Lasak I, Metspalu M, Moiseyev V, Gromov A, Pokutta D, Saag L, Varul L, Yepiskoposyan L, Sicheritz-Pontén T, Foley RA, Lahr MM, Nielsen R, Kristiansen K, Willerslev E - Cell (2015)

Authenticity of Y. pestis DNA(A) DNA damage patterns for RISE505 and RISE509. The frequencies of all possible mismatches observed between the Y. pestis CO92 chromosome and the reads are reported in gray as a function of distance from 5′ (left panel, first 25 nucleotides sequenced) and distance to 3′ (right panel, last 25 nucleotides). The typical DNA damage mutations C>T (5′) and G>A (3′) are reported in red and blue, respectively.(B) Ancient DNA damage patterns (n = 7) of the reads aligned to the CO92 chromosome and the Y. pestis associated plasmids pMT1, pCD1 and pPCP1. The boxplots show the distribution of C-T damage in the 5′ of the reads. The lower and upper hinges of the boxes correspond to the 25th and 75th percentiles, the whiskers represent the 1.5 inter-quartile range (IQR) extending from the hinges, and the dots represent outliers from these.(C) DNA fragment length distributions from RISE505 and RISE509 samples representing both the Y. pestis DNA and the DNA of the human host. The declining part of the distributions is fitted to an exponential model (red).(D) Linear correlation (red) between the decay constant in the DNA of the human host and the associated Y. pestis DNA extracted from the same individual (R2 = 0.55, p = 0.055). The decay constant (λ) describes the damage fraction (i.e., the fraction of broken bonds on the DNA strand).(E) Distribution of edit distance of high quality reads from RISE505 and RISE509 samples mapped to either Y. pestis (dark gray) or Y. pseudotuberculosis (light gray) reference genomes. The reads have a higher affinity to Y. pestis than to Y. pseudotuberculosis.(F) Plots of actual coverage versus expected coverage for the 101 screened samples. Expected coverage was computed taking into account read length distributions, mappable fractions of reference sequences, and the deletions in pMT1 for some of the samples. Samples assumed to contain Y. pestis are shown in blue and RISE392 that is classified as not Y. pestis appears is shown in red. See also Figure S1 and S2, Table S3.
© Copyright Policy - CC BY
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4644222&req=5

fig3: Authenticity of Y. pestis DNA(A) DNA damage patterns for RISE505 and RISE509. The frequencies of all possible mismatches observed between the Y. pestis CO92 chromosome and the reads are reported in gray as a function of distance from 5′ (left panel, first 25 nucleotides sequenced) and distance to 3′ (right panel, last 25 nucleotides). The typical DNA damage mutations C>T (5′) and G>A (3′) are reported in red and blue, respectively.(B) Ancient DNA damage patterns (n = 7) of the reads aligned to the CO92 chromosome and the Y. pestis associated plasmids pMT1, pCD1 and pPCP1. The boxplots show the distribution of C-T damage in the 5′ of the reads. The lower and upper hinges of the boxes correspond to the 25th and 75th percentiles, the whiskers represent the 1.5 inter-quartile range (IQR) extending from the hinges, and the dots represent outliers from these.(C) DNA fragment length distributions from RISE505 and RISE509 samples representing both the Y. pestis DNA and the DNA of the human host. The declining part of the distributions is fitted to an exponential model (red).(D) Linear correlation (red) between the decay constant in the DNA of the human host and the associated Y. pestis DNA extracted from the same individual (R2 = 0.55, p = 0.055). The decay constant (λ) describes the damage fraction (i.e., the fraction of broken bonds on the DNA strand).(E) Distribution of edit distance of high quality reads from RISE505 and RISE509 samples mapped to either Y. pestis (dark gray) or Y. pseudotuberculosis (light gray) reference genomes. The reads have a higher affinity to Y. pestis than to Y. pseudotuberculosis.(F) Plots of actual coverage versus expected coverage for the 101 screened samples. Expected coverage was computed taking into account read length distributions, mappable fractions of reference sequences, and the deletions in pMT1 for some of the samples. Samples assumed to contain Y. pestis are shown in blue and RISE392 that is classified as not Y. pestis appears is shown in red. See also Figure S1 and S2, Table S3.
Mentions: Besides applying standard precautions for working with ancient DNA (Willerslev and Cooper, 2005), the authenticity of our findings are supported by the following observations: (1) The Y. pestis sequences were identified in significant amounts in shotgun data from eight of 101 samples, showing that this finding is not due to a ubiquitous contaminant in our lab or in the reagents. Indeed, further analysis showed that one of these eight was most likely not Y. pestis. We also sequenced all negative DNA extraction controls and found no signs of Y. pestis DNA in these (Table S3). (2) Consistent with an ancient origin, the Y. pestis reads were highly fragmented, with average read lengths of 43–65 bp (Table S3) and also displayed clear signs of C-T deamination damage at the 5′ termini typical of ancient DNA (Figure 3, Figure S1). Because the plasmids are central for discriminating between Y. pestis and Y. pseudotuberculosis, we tested separately for DNA damage patterns for the chromosome and for each of the plasmids. For the seven samples, we observe similar patterns of DNA damage for chromosome and plasmid sequences (Figure 3, Figure S1). (3) We observe correlated DNA degradation patterns when comparing DNA degradation in the Y. pestis sequences and the human sequences from the host individual. Given that DNA decay can be described as a rate process (Allentoft et al., 2012), this suggests that the DNA molecules of the pathogen and the human host have a similar age (Figure 3, Figure S1, Table S3 and Supplemental Experimental Procedures). (4) Because of the high sequence similarity between Y. pestis and Y. pseudotuberculosis, we mapped all reads both to the Y. pestis CO92 and to the Y. pseudotuberculosis IP32953 reference genomes (Chain et al., 2004). Consistent with being Y. pestis, the seven investigated samples displayed more reads matching perfectly (edit distance = 0) toward Y. pestis (Figure 3, Figure S2). One sample (RISE392) was most likely not Y. pestis based on this criterion. (5) A naive Bayesian classifier trained on known genomes predicts the seven samples to be Y. pestis with 100% posterior probability, while RISE392 is predicted to have 0% probability of being Y. pestis (Figure S2, Table S3). (6) If the DNA was from other organisms than Y. pestis, we would expect the reads to be more frequently associated with either highly conserved or low-complexity regions. However, we find the reads to be distributed across the entire genome (Figure S2), and comparison of actual coverage versus the coverage that would be expected from read length distributions and mappability of the reference sequences are also in agreement for the seven samples (Figure 3). (7) In a maximum likelihood phylogeny, the recovered Y. pestis genomic sequences of RISE505 and RISE509 are clearly within the Y. pestis clade and basal to all contemporary Y. pestis strains (Figure 4) (see below).

Bottom Line: How and when it originated remains contentious.We also identify a temporal sequence of genetic changes that lead to increased virulence and the emergence of the bubonic plague.Our results show that plague infection was endemic in the human populations of Eurasia at least 3,000 years before any historical recordings of pandemics.

View Article: PubMed Central - PubMed

Affiliation: Center for Biological Sequence Analysis, Department of Systems Biology, Technical University of Denmark, Kemitorvet, Building 208, 2800 Kongens Lyngby, Denmark.

Show MeSH
Related in: MedlinePlus