iiinova — where your digital self keeps evolving

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

7) With our refined phylochronology, we decided to reassess all published genomes along with our new ones, resulting in a 60-page SI section, providing a critical discussion of all data, our reasoning, and primary sources supporting the suggested plague outbreaks associated with the genomes.

Table 2 of the paper showing previously published genomes and the historically recorded outbreaks they are putatively linked to.

021

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

6) We identify several clusters of identical genomes, i.e. incidents in which strains spread widely without accumulating additional mutations. One of those events (N18) happened in the Baltic region right after the Little Polytomy (N16), highlighting this period as particularly proliferative.

Cutout of the schematic tree highlighting the Little Polytomy and the closely related cluster of identical Baltic genomes.

100

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

5) Working with non-UDG data is challenging, since environmental contamination cannot be easily filtered while allowing for deamination damage. We developed 2 new approaches to reduce the # of false SNPs in those genomes without affecting phylogenetically informative positions.

Plots showing the number of private SNPs of non-UDG-derived genomes after different treatments as well as pairwise statistics.

100

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

4) In the absence of sufficient RC metadata, we applied a generic offset and got satisfactory results that we could cross-validate against other chronological data. Despite the increased uncertainty through applying a generic offset, our model was able to significantly shorten the RC intervals.

Figure showing the plot for the main radiocarbon model and box plots showing the distribution of radiocarbon spans and intervals before and after modeling.

100

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

3) Modeling with uncorrected radiocarbon dates results in poor support since several RC dates appear too old. We conclude that RC dates of human remains are often affected by reservoir effects and offsets that need to be considered in contexts where a difference of a decade changes the whole story.

Clipping from a radiocarbon modeling plot showing low agreement indices for two radiocarbon dates.

100

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

2) We describe a Little Polytomy around 1450-1500 that gave rise to two major clades and – to our current knowledge – additional short-lived clades. Another lineage surviving for at least a century branching just one SNP before, suggesting that several new reservoirs were seeded during this period.

Phylogenetic and schematic tree of the Second Pandemic genomes.

110

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

Some highlights: 1) We identify repeated introductions of plague into Estonia, from the late 14th c. until one of the last major outbreaks in 1710 during the Great Northern War. We found plague victims at both urban and rural sites, showing how well-connected the region was to the rest of Europe.

Map showing the location of all plague genomes of the Second Pandemic.

120

marcelkeller.bsky.social @marcelkeller.bsky.social · 11/09/2026

Our paper on the second plague pandemic is now finally published: www.pnas.org/doi/10.1073/... We present 11 new Y. pestis genomes from around Europe and introduce a new approach to refine the dating of ancient genomes that helped us to contextualize 75 plague genomes of the Second Pandemic.

Schematic workflow showing the Phylogenetically Informed Radiocarbon Modeling

1198