Genomic windows into ancient epidemics

Population genomic analyses of ancient and modern individuals are providing unprecedented resolution in our view of past human demography and adaptive episodes. Here, I highlight three recent studies in the field that investigated ancient human encounters with major pathogens.

Like Comment

Living through a global pandemic may only recently have brought infectious disease, immunity, and evolution into sharp public focus, but interactions with pathogens have been shaping our biology through our entire evolutionary history. Pathogens that affect substantial proportions of the host population often leave recognizable signatures of selection in host genomes. An early demonstration of how pathogens have left impressions on human genomes and biology came through the discovery of high frequencies of the sickle cell-causing genetic variant in African populations in which malaria has been historically prevalent, as the trait confers protection against the malarial parasite.

Advances in whole-genome sequencing over the last two decades have yielded a plethora of studies revealing broad patterns of pathogen-driven evolution in humans. These studies have ranged from genome-wide scans for positive selection that showed rapid evolution  of many immune genes, to investigations of the impact of agriculture, and resulting high population densities and close animal contact on immunity.

We are perhaps approaching a new frontier in the study of our relationships with many major pathogens. Analysis of ancient DNA from archaeological remains of individuals who lived at various points in the past has emerged as a particularly powerful tool, providing a direct look at patterns of genetic diversity in the past. These genomic snapshots can not only resolve questions about the origins and spread of modern human populations, but also help address or generate hypotheses about historical events, such as pandemics. At the same time, methodological advances in reconstructing evolutionary history from modern genomes allow increasingly precise dating of past adaptive events. Three recent studies provide roadmaps for learning about the nature and evolutionary impact of specific episodes of past human interactions with known pathogens.

Autopsying the Black Death

The Yersinia pestis-driven Black Death pandemic in 1346-1353, along with a series of outbreaks that followed over the next few centuries, resulted in millions of deaths across Europe. It has long been thought that the high mortality rate of the disease likely impacted the frequencies of genetic variants (or alleles) that conferred protection against, or susceptibility to Y. pestis in European populations. While several studies have identified putative plague-selected variants through population genetic inference based on modern genetic variation, a recent study in Molecular Biology & Evolution took a more direct approach by examining genetic variation across almost 500 immune-related genes in 36 individuals buried in three 16th century plague mass graves in Ellwangen, Germany.

Comparison of these ancient individuals with modern individuals native to the same region revealed changes in frequencies of variants in several of the examined genes, including FCN2, which encodes ficolin-2, a protein involved in innate immunity through the recognition of various components of the bacterial cell wall. Variant frequencies tend to fluctuate in populations just by chance in a process known as genetic drift, and not all changes in frequencies are brought about by the action of natural selection. To exclude such random changes in this instance, the authors performed thousands of simulations with no selection acting on the variants to show that the observed changes were very unlikely to have occurred by chance and thus probably driven by selection. They did not find evidence of selection acting on many of the putative variants identified in earlier studies, but it is hard to rule them out based on analysis of individuals from a single region.

A history of consumption

While the above study took advantage of historical information to help identify and contextualize changes in the frequencies of genetic variants, another recent paper, published in the American Journal of Human Genetics, was able to use changes in frequency of a well-characterized variant to reconstruct history. Tuberculosis (TB) – caused by the bacterium Mycobacterium tuberculosis – has deep roots in Europe, known to Ancient Greeks as phthisis and its symptom of weight loss earning it the name of consumption in 17th century England. The authors sought to investigate the historical burden of the disease based on a genetic proxy.

A single mutation in the gene TYK2, which encodes a protein involved in immune signaling, increases the risk of experiencing severe outcomes from TB in individuals who carry the mutation in both of their gene copies, and would therefore be expected to be selected against when TB is prevalent. Genomic data from more than a thousand individuals from archaeological sites spanning the Holocene and beyond in Europe enabled the authors to directly follow the trajectory of the frequency of the TYK2 variant over time. They found that the variant had been maintained by a combination of random genetic drift and genetic exchange between subpopulations for thousands of years, until its frequency underwent a steep decline starting from about 2,000 years ago.

A simplified depiction of Kerner et al.'s premise. Illustration by author; map outline from

The authors simulated variant frequency changes under several regimes of negative selection or no selection to show that the observed decline was likely driven by negative selection acting since about 2,000 years ago. This implies a high TB burden in European populations over the last two millennia, which selected against the TYK2 variant conferring susceptibility to severe TB.

A distant harbinger

Even as the ancient DNA revolution continues to provide firsthand genomic accounts of the past, we are not quite done mining all the evolutionary information stored in the much more abundant genomic data from modern human populations. One such recent preprint study is of particular interest due to the pathogen(s) being considered. While our interactions with SARS-CoV-2 – the virus responsible for the ongoing pandemic – represent a mere flirtation over evolutionary timescales, it is certainly not the first member of the coronavirus family that we have encountered. Within most of our lifespans, there have been the MERS and original SARS outbreaks, but the preprint takes us back much further.

The study investigated genes encoding proteins that are known to interact with, and possibly help defend against modern coronaviruses. As genomes pass through generations, recombination events between parental chromosomes split apart adjacent genetic regions at a certain rate. When a beneficial genetic variant arises, it typically spreads through the population faster than recombination can separate it from its neighboring regions, leading to reduced genetic diversity surrounding the variant. Using methods that exploit this genomic signature, the authors found evidence that many variants occurring in the coronavirus-interacting genes had spread rapidly due to positive selection in multiple East Asian populations, but not in other examined populations from the rest of the world. This implies ancient encounters with coronavirus(es), or other virus(es) with very similar biology, but when?

To answer this question, the authors employed a recently developed method that can reconstruct the genealogical histories of individual variants based on their frequencies in modern populations, patterns of diversity in neighboring genomic regions, and the rate at which mutations accumulate over time in the human genome. Positive selection appears to have acted on many of the examined variants over an extended period between about 20,000 and 5,000 years ago, indicating a long-term presence and burden of coronavirus-like infections in East Asia.

Could the ongoing pandemic leave genomic signatures of selection that are detectable thousands of years in the future? And would genomic adaptations against the virus be desirable? Consider for a moment the cost of selection. A genomic signature of positive selection in the descendants of survivors of past epidemics simultaneously signifies the death, or at the very least lower reproductive success of individuals containing competing genetic variants. Similarly, a decline over generations in the frequency of a disease susceptibility variant due to negative selection reflects the loss of individuals carrying the variant, again through death or lowered fertility. For modern humans to still retain detectable genomic signatures of selection from these ancient encounters with pathogens as we see in these studies, humans of the past likely incurred costs that we can and should no longer accept. Armed with modern public health measures and medical technology, we should certainly hope to never have to biologically adapt to pathogens again.

Primary references

Immel, A. et al. Analysis of genomic DNA from medieval plague victims suggests long-term effect of Yersinia pestis on human immunity genes. Molecular Biology and Evolution (2021) doi:10.1093/molbev/msab147.

Kerner, G. et al. Human ancient DNA analyses reveal the high burden of tuberculosis in Europeans over the last 2,000 years. The American Journal of Human Genetics 108, 517–524 (2021).

Souilmi, Y. et al. An ancient viral epidemic involving host coronavirus interacting genes more than 20,000 years ago in East Asia. bioRxiv 2020.11.16.385401 (2021) doi:10.1101/2020.11.16.385401.

Ornob Alam

PhD candidate, New York University

I am a PhD candidate in Michael Purugganan's lab at NYU, where I study the evolutionary and demographic history of domesticated Asian rice.