Another look at selection and the Black Death

The immune system is a crucial biological interface between humans and our environments. This dimension of human evolution has long been of interest to human geneticists, through analysis of variation of HLA, blood types, and other polymorphisms.

Immunity has become accessible to study directly in ancient groups only recently, with the advent of ancient DNA sequencing. As more and more evidence emerges, understanding ancient immune dynamics is changing our views of how populations evolved.

In my top 10 stories about ancient people from DNA from 2022, I recognized work by Jennifer Klunk, Tauras Vilgalys, and coworkers on evidence for selection associated with the Black Death. Their study considered skeletal remains of people in the East Smithfield mass grave in central London, who died during the plague outbreak of 1346–1352 CE, compared to cemetery samples from before and after the outbreak. The study also included cemetery samples from Denmark before and after the plague. They hypothesized that a genetic marker associated with plague survival would show a pattern of reduced frequency in plague deaths and higher frequency in survivors, who died later. They found a number of markers fitting this pattern.

One of these candidates is a single nucleotide polymorphism linked to the gene ERAP2. Because this gene was already known for its roles in immunity, Klunk and coworkers designed experiments to assess whether the hypothesized protective allele had any effects on Yersinia pestis. What I found to be the most interesting part of the study is that the gene did indeed matter to immune cell activity related to plague.

It's really rare to get this kind of cellular or functional evidence in association with population genetic signs of selection. When I began working on genetic evidence of natural selection, back in the early 2000s, I was often encountering population genetic evidence for selection on gene regions that had no known function. Even in the best-known cases of selection, like lactase persistence, where the population genetic evidence for selection was very strong and the proximity of the selected SNPs to the LCT gene was very persuasive, nonetheless it took a long time to develop functional evidence of how the particular allelic differences mattered to lactose expression. While this was being worked out, I talked to a number of human geneticists who doubted that lactose digestion was important enough to ancient people to have been subject to natural selection!

That note of background may help to explain why I think this kind of work is valuable.

Still, the sample size of ancient genomes from plague cemeteries is not that big. Statistical power for tests of allele frequency change is low for small samples, and the chance of seeing a big frequency change for a SNP just due to sampling error is substantial. Correcting for multiple comparisons is essential and somewhat tricky

This week a challenge to the ERAP2 plague work was published in Nature written by Alison Barton and several coauthors who do a lot of work with ancient genomes. This comment points out some weaknesses in the population genetic side of the plague selection argument. The sample is too small for reliable inference of selection, there seem to be batch effects from different genotyping approaches in the various samples that may have given rise to spurious frequency differences, and outside the context of these samples there is no further evidence for directional selection on ERAP2.

In a reply to this comment, Vilgalys, Klunk, and coworkers present some reanalysis of their population genetic data to answer the issues raised by Barton and coworkers. They support their original argument for selection on ERAP2, and reiterate the value of the functional evidence that they gathered.

They also bring attention to another aspect of their earlier study that I had neglected in 2022. The allelic variation of ERAP2 is quite old in the human population, and some previous work had shown statistical evidence of balancing selection maintaining that variation. The gene is not an example of long-term directional selection, like lactase persistence. Instead, they propose a dynamic something like HLA allelic variation in which short episodes where one or another allele have a selective advantage balance out over long evolutionary time.

A pattern of occasional short bursts of directional selection as part of a longer-term pattern of balancing selection was probably very common among immune-related genes across ancient human populations. This is one of the complicated parts of trying to understand natural selection in past populations; as a result of environments, natural selection changed in pattern and intensity when environments changed. Few environmental factors change on a dime like pathogens.

All that is to say I think that Vilgalys, Klunk, and coworkers have got some interesting results and have sketched out a productive way of investigating selection associated with epidemics. That approach would of course benefit from larger samples and more geographic locations to confirm the pattern they found.

When I started work in human evolution, one of my first lines of investigation was the biological changes that unfolded with agriculture and village life. Pathogen dynamics were long understood to have changed massively with these shifts in human cultures. While it made eminent sense that many dimensions of immunity would have been under strong selection across the Holocene, the data to investigate this would take a while to get going. I had little guess that we would be able to see such information directly from ancient genomes.

What has been exciting in the meantime is seeing that immunity underwent many other transitions in our evolutionary history. The introduction of Neandertal and Denisovan immune genes into modern human populations gave a massive clue that evolution of immunity has not just followed our recent environmental changes but may have been the strongest driver of our evolution for hundreds of thousands of years.

References

Andrés, A. M., Dennis, M. Y., Kretzschmar, W. W., Cannons, J. L., Lee-Lin, S.-Q., Hurle, B., Program, N. C. S., Schwartzberg, P. L., Williamson, S. H., Bustamante, C. D., Nielsen, R., Clark, A. G., & Green, E. D. (2010). Balancing Selection Maintains a Form of ERAP2 that Undergoes Nonsense-Mediated Decay and Affects Antigen Presentation. PLOS Genetics, 6(10), e1001157. https://doi.org/10.1371/journal.pgen.1001157

Barton, A. R., Santander, C. G., Skoglund, P., Moltke, I., Reich, D., & Mathieson, I. (2025). Insufficient evidence for natural selection associated with the Black Death. Nature, 638(8051), E19–E22. https://doi.org/10.1038/s41586-024-08496-5

Klunk, J., Vilgalys, T. P., Demeure, C. E., Cheng, X., Shiratori, M., Madej, J., Beau, R., Elli, D., Patino, M. I., Redfern, R., DeWitte, S. N., Gamble, J. A., Boldsen, J. L., Carmichael, A., Varlik, N., Eaton, K., Grenier, J.-C., Golding, G. B., Devault, A., … Barreiro, L. B. (2022). Evolution of immune genes is associated with the Black Death. Nature, 611(7935), Article 7935. https://doi.org/10.1038/s41586-022-05349-x

Vilgalys, T. P., Klunk, J., Demeure, C. E., Cheng, X., Shiratori, M., Madej, J., Beau, R., Elli, D., Patino, M. I., Redfern, R., DeWitte, S. N., Gamble, J. A., Boldsen, J. L., Carmichael, A., Varlik, N., Eaton, K., Grenier, J.-C., Golding, G. B., Devault, A., … Barreiro, L. B. (2025). Reply to: Insufficient evidence for natural selection associated with the Black Death. Nature, 638(8051), E23–E29. https://doi.org/10.1038/s41586-024-08497-4