Apaper published two days ago, on Jan 22, 2026, reported that the authors had detected and dated the genome of a bacteria belonging to the Treponema genus, which causes syphilis, yaws, bejel, and pinta. The pathogen that was found in human remains that are 5,500 years old, in Bogotá, Colombia.
The paper published in Science is the following: Davide Bozzi et al.(2026). A 5500-year-old Treponema pallidum genome from Sabana de Bogotá, Colombia. Science 391, eadw3020. DOI:10.1126/science.adw3020
The microbe is a Treponema pallidum subspecies, but it is closer to the base of the tree, an older relative of the other branches, that carry subspecies that cause syphilis, yaws, pinta, and bejel.
The paper then tried to date how old this variant, named TE1-3, is. They estimated it split around 13,700 years ago from its sister lineatges: 6768 to 20,592 cal yr B.P. The authors noted that "This suggests that the divergence between TE1-3 and the modern, genomically characterized T. pallidum subspecies occurred during the Late Pleistocene to Early Holocene, closely following the peopling of the Americas, whereas the diversification of the known subspecies themselves took place more recently, within the Holocene ~6000 cal yr B.P. (3622 to 9452 cal yr B.P. 95% HPD)."
As mentioned in previous posts, I wonder if the date fit was "eased" into a recent window that coincides with the currently accepted date for the peopling of America. What would a date of 40,000 y BP have meant?. The Methods don't specify how this date was calculated. The methods describes the process:
"Given TE1-3 is several thousand years older than all ancient T. pallidum sequenced to date, it is an ideal tip calibration for molecular clock analyses. To assess temporal signal, we performed a root-to-tip regression using radiocarbon-dated ancient genomes and collection dates for modern genomes (table S9). This analysis confirmed a strong temporal structure (R2 = 0.67; fig. S17).
We then applied tip-dated Bayesian inference in BEAST2 (49), which yielded a mean substitution rate of 8.86 × 10−8 substitutions per site per year [95% Highest Posterior Density (HPD): 6.5 to 11.2 × 10−8]. The divergence of TE1-3 from the other T. pallidum lineages is estimated at 13,746 cal yr B.P. (95% HPD: 6768 to 20,592 cal yr B.P. )"
However, the root-to-tip regression used a lot of recent data points from the late 1900s, and 2010s a few from the 1700s, and 1500s, and then a few older ones from 1371 AD, 1329 AD, 3438 BP. These are given in their Table S9 (Supplementary table S9. Tip dates used in assessment of temporal signal and root-to-tip regression).
"Regression" is a statistical tool that helps see the trend in a series of scattared data points, assign a function to explain the data and use that function to calculate other data points. For instance, the black dots in the image below are data points and the red line is the best fitting curve (in this case it is a linear regression of the y= a + bx type, the equation for a line. Once the line is known, we can extrapolate the value of y for a given x value.
The fitting isn'd done manually, there is a mathematical-statistical tool to do so. The so called "correlation coefficient" or R squared (R2 mentioned above) its value ranges from 0 to 1, where 0 is no correlation and 1 is a 100% correlation. In this case they state the value is 0.67. But in my engineer point of view, it is heavily skewed by the data points in their Figure S17 shown below, where all the data points (upper left) are lined in vertical strips, and one solitary data point dated 3438 BP in the lower left is used to align the function.
Furthermore, in their Fig. 17, they report two lower values for R2, of 0.294 and 0.127!
The authors caption for Fig. S17. "Root-to-tip regression. Inspection of the temporal signal of the dataset... the root-to-tip regression displaying the temporal signal of the dataset (black line) and of specific clades: TPA in dark green and TPE/TEN in light green. For the global clock rate we observe a rate of 1.336 x 10^-7 with an R2 of 0.667, while a weaker temporal signal (lower R2) but similar evolutionary rates were observed within the other two clades: 1.202 x 10^-7 for TPA (R2 = 0.127) and 1.190 x 10^-7 for TPE/TEN (R2 = 0.294). Plot generated with clockor2 (112)".
The authors also state that " We see that using the even more divergent Cuniculi A genome as a reference pushes the date back in time. Consequently, our estimate when using the Nichols reference genome is best viewed as a lower bound for divergence of TE1-3." It seems that the split between Old World human variants and the Amerindian one would have been older if they used the Treponema paraluisleporidarum Cuniculi A genes as a basal reference! The paper mentions how the "trimmed" genes to adapt it to their needs.
Comments
The paper is relevant because it gives an American origin, and an old date of over 13 ky, to a bacteria with high virulence: "All previously identified virulence factors common to Treponema pallidum are present in TE1-3, including immune evasion and tissue tropism tpr paralogs, suggesting that it possesses the same virulence potential as T. pallidum."
Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2026 by Austin Whittall ©
