Ancient DNA as a tool for medical research
The simplest model used to study human susceptibility to infection is genetic predisposition to infectious diseases through inborn errors of immunity — mutations that increase the risk of severe infections6. The most successful approach has been to study people who are extremely susceptible to infections, as they have the highest odds of carrying highly penetrant genetic lesions. The main advantage of this approach is that causality between genotype and phenotype follows naturally from the study of an in vivo (human) model, although its implementation requires the extensive genetic screening of severely ill patients.
An alternative approach is to study the effects of natural selection from pathogenic pressure on human genome variability7. These two frameworks are comparable in that both involve the identification of variants that increase the risk of a given infectious disease in natura. However, they differ in that inborn errors of immunity usually operate at the scale of a single generation, whereas gradual pathogenic pressure operates at the scale of many generations, which makes it possible to identify genetic variants with effect sizes that differ by several orders of magnitude.
Recent studies have highlighted the value of using ancient genomes from different epochs, known as aDNA time series, to reconstruct the evolutionary history of immune disorders and past epidemics (Fig. 1). One recent proof-of-concept study of more than 1,000 genomes dated to within the last 10,000 years of European history showed that a tuberculosis risk variant, TYK2 P1104A, present in around 3% of people of European ancestry, has evolved under strong negative selection over the past two millennia8. This finding, probably reflective of the pressure imposed by Mycobacterium tuberculosis, would have been very difficult to achieve through studies of modern DNA. Indeed, most methods for detecting natural selection in modern DNA data are underpowered for low-frequency variants.
Schematic representation of the use of aDNA to study the effects of negative selection on genetic variants associated with disease risk in the context of past pathogenic pressure. Here, the aDNA samples date from either before or after an epidemic event. The frequencies of DNA variants between the two groups of samples (pre- and post-epidemic groups) can identify genetic variants that are targeted by negative selection, which are present at a significantly lower frequency than would be expected by chance in the post-epidemic group. Such observations provide clues to the pathogenic nature of the variant, and the corresponding gene, in the context of the infectious disease studied.
The evolutionary history of the pathogen itself can also provide insight into the dynamics of past epidemics, but this was difficult to characterize until samples from ancient pathogens became available. In the context of tuberculosis, studies of ancient mycobacteria have dated the most recent common ancestor of the pathogen to only 6,000 years ago, which contrasts markedly with the estimates of more than 70,000 years ago that were obtained from studies of modern strains of M. tuberculosis9. Medical practitioners may find aDNA studies similarly helpful for pinpointing genetic variants of microorganisms, as their evolutionary history can reveal their deleteriousness to human health.
aDNA time series at the scale of the entire human genome can identify variants under negative selection, which is useful for the detection of new genetic factors associated with immune disorders, as described in a recent study10. Paleogenomics therefore appears to be a powerful approach, complementary to epidemiological and clinical genetics studies, that can be used to confirm and expand on genetic variants associated with disease risk.
