Ancient DNA (aDNA) has helped us to understand past human diets by providing an insight into the relationship between dietary habits and their impact upon the genome. To review this further, studies on lactose persistence and starch consumption have been selected. These studies will be placed in the context of Niche Construction Theory (NCT), providing a more critical understanding of the diet-genetic relationship. However, more work can be done to reach the full potential of this type of analysis. Furthermore, there are areas of past diets and their relationship to our genome which have not been fully understood. These areas may provide a basis for future research such as celiac disease.
1 Niche Construction Theory
Niche Construction Theory (NCT) recognises individual agency as a variable in the evolution of organisms (O’Brien and Bentley 2020, 1). NCT suggests that organisms can alter natural selection through instigating cultural change in their environment, which consequently begins to impact the genome of the organisms in the ecosystem. This means that organisms act as ‘co-editors’ of their own and other species’ evolution (Laland and O’Brien 2010, 303). Examples of culturally-included changes include the domestication of livestock, starting fires, and building shelters (Laland and O’Brien 2010, 307). Genetic changes in humans have also occurred through dietary changes. As diet is often associated with an individual culture, it can therefore be assumed that these dietary changes are a culturally-induced change. By using NCT and the genetic-culture relationship as a theoretical framework, how far aDNA analysis has contributed to our understanding of past diets can be further understood.
2 Lactose: theoretical perspectives
Most mammals cannot digest products containing high amounts of lactose into adulthood (Feldman and Cavalli-Sforza 1989, 146). However, some adult humans can digest lactose into adulthood without any side-effects; they are lactase persistent (LP). Individuals who cannot digest lactose are referred to as non-lactase persistent (NLP), and effects include diarrhoea, bloating, and chronic flatulence (Leonardi et al. 2012, 89). LP individuals are often found in societies which have a long-standing association with cattle. Based upon this, two main theoretical models on the evolution of LP have been produced.
The first model is the Reverse Culture Hypothesis (RCH) (See Figure 1). RCH suggests that societies with a long-standing link to cattle are lactose persistent because lactose consumption was adopted by cultures who were inherently lactose tolerant (McCracken 1971, 495). Those who consumed dairy and found themselves lactose intolerant rejected dairy due to the side-effects (McCracken 1971, 495). Therefore, LP cultures had a higher frequency of LP alleles before the uptake of dairying when compared to non-lactose persistent cultures.
In contrast, the Culture Historical Hypothesis (CHH) (see Figure 2) suggests that there was a variation in the frequency of LP and NLP alleles before the uptake of dairying in all cultures. The genomes of cultures adopting dairying underwent a strong selective pressure for LP due to LP individuals having a nutritional advantage (Feldman and Cavalli-Sforza 1989, 146). These circumstances exerted a strong positive selective pressure on the LP alleles, increasing the frequency of LP alleles over NLP alleles in the population. Overall, this model also engages with Niche Construction Theory as it suggests that the human genome has been changed through the cultural change of dairying.
One way in which aDNA has contributed to our understanding of past human diets is by providing evidence in support of the Culture Historical Hypothesis and NCT over the Reverse Culture Hypothesis, developing our understanding of the relationship between the human genome and past diets further.
Burger et al. (2007) display evidence in support of the Culture Historical Hypothesis and Niche Construction Theory by providing a temporal basis for the change in the frequencies of LP and NLP alleles. This suggested that there was a variation of LP and NLP frequencies within cultures before the uptake. Eight skeletons were selected from the European central northeast Neolithic, the southeast Neolithic, the Mesolithic, and the Medieval period (see Table 1) (Burger et al. 2007, 3736). The samples were analysed for the identification of the European LP variant 13.010*T. PCR amplification, capillary electrophoresis and oligonucleotide primer pairs were used for SNP analysis and aDNA authentication (Burger et al. 2007, 3736). The presence of the LP variant was identified in the Neolithic and Medieval samples, with an absence in the Mesolithic sample. The authors suggested that this was due to the Neolithic and Medieval samples originating from cultures which practiced dairying, therefore providing evidence for a selective sweep for LP in favour of the Culture Historical Hypothesis (Burger et al. 2007, 3740). Overall, this type of aDNA analysis has contributed to our understanding of past diets as this study provided evidence for the CHH. However, the major contribution of this work can be seen in the context of Niche Construction Theory. As the results have provided evidence for a selective sweep after the introduction of dairying, this shows how an alteration in the environment due to dairying can exert a selective pressure on the human genome, providing scientific evidence for the relationship between the past human diet and the human genome.
A second study which represents how aDNA analysis has contributed to our understanding of past human diets is by Malmström et al. (2010). Fourteen Neolithic samples were selected from sites in Gotland and the Baltic Sea (Malmström et al. 2010, 2; See Figure 3). All of the samples were at least 1,000 years younger than the advent of agriculture in Scandinavia, and therefore gave an insight into LP allele frequencies before dairying. Primers were used to target C/T polymorphisms with pyrosequencing for allele identification (Malmström et al. 2010, 2-3). Furthermore, computer modelling was used to suggest that the frequency of the 1310*T allele (LP) was low in the middle Neolithic hunter-gatherer population when compared to the extant Swedish population (Malmström et al. 2010, 1). The results suggested that the LP variant 1310*T has undergone positive selective pressure between the middle Neolithic and present day. This strong selective pressure did not exist before the middle Neolithic, and therefore the aDNA analysis supports the Cultural Historical Hypothesis. How far this work has shown that aDNA analysis has contributed to our understanding of past human diets can also be measured by placing this work in the context of NCT. Between the Mesolithic and Middle Neolithic, the only dietary change that occurred in relation to lactose persistence would have been the increased consumption of lactose products, resulting from the cultural and environmental change of the uptake of dairying. Therefore, this work provides evidence of humans inadvertently acting as ‘co-editors’ of their own genomes by creating a positive selective pressure for LP. As argued by Burger et al. (2007); aDNA has contributed hugely to our understanding of past diets by providing a glimpse into the complicated relationship between past dietary changes and the subsequent evolutionary pressures placed upon the human genome.
Overall, this analysis has contributed to our understanding of past diets by providing evidence in support of the Culture Historical Hypothesis. Throughout, it was argued that the overarching contribution of aDNA analysis to our understanding of past diets is a further insight into the relationship between a culturally-induced dietary change and the selective impact upon evolution of our genome, as suggested by Niche Construction Theory.
2.2 Lactose: aDNA potential
Whilst the above studies have provided some evidence of the frequencies of LP variants in a few prehistoric samples, these samples are not representative of all past populations. aDNA analysis into samples which are more representative of a societal group may provide more insight. Massacre sites in the Neolithic are found at the end of the LBK period. As deaths in a massacre occur in a short period of time, the temporal resolution of the samples is high and therefore representative of a living population (Bentley and O’Brien 2019, 99). Candidates for the further investigation of lactase persistence using aDNA analysis are the sites of Talheim and Herxheim. At Talheim, the remains of thirty-four individuals were recovered, all buried in a single event (Whittle and Bickle 2013, 486). At Herxheim, up to five-hundred individuals have been recovered (Wild et al. 2004, 380). At each site, all age ranges and each biological sex are represented. Therefore, these sites may improve our understanding of past human diets through aDNA analysis by providing a more representative insight into LP variant frequencies within agrarian societies in the Neolithic.
A second way in which aDNA has contributed to our understanding of past diets is the evolution of the amylase gene, which can also be analysed through the lens of NCT. Amylase is the enzyme that digests starch and is expressed in the saliva due to the AMY1 gene (Marciniak and Perry 2017, 666). It is possible that starch consumption is a key trait of agricultural societies and hunter-gatherers in arid environments; and individuals with high-starch diets, on average, have more AMY1 copies (Perry et al. 2007, 1256). A high copy number of AMY1 is thought to be beneficial as the expression of salivary amylase is increased (Inchley et al. 2016, 1). This has been the basis for some debate, with some suggesting that selection may have begun in the Neolithic transition due to an increase in starch-rich domesticated crops (Perry 2007, 1256). However, others have suggested that the past human dietary change and selective pressure may have occurred before the Neolithic transition with the advent of cooking (Inchley et al. 2016). This is because amylase is poor at digesting starch if the starch has not first been gelatinised by heat (Carmody et al. 2016, 1092). Evidence of the consistent use of hearths can be found from c.300,000-years-ago (Shahack-Gross et al. 2014), and archaeological evidence suggests that the controlled use of fire occurred in the Middle Palaeolithic (Gowlett and Wrangham 2013).
3.1 Starch: aDNA contributions
The first starch-based case study is the work of Lazaridis et al. (2014), which provides evidence of the timing of a selective pressure for an increased AMY1 copy number. The team sequenced the genomes of a c.7,000-year-old farmer from Germany, alongside eight c.8,000-year-old hunter-gatherers from Luxembourg and Sweden (Lazaridis et al. 2014, 490; see Figure 4). The aDNA was authenticated by identifying deamination-derived mismatches at the ends of molecules (Lazaridis et al. 2014, 409). Analysis found that the Stuttgart Neolithic farmer had the highest AMY1 copy number, suggesting a consistency in the relationship between a high-starch diet due to the introduction of agriculture and an increased copy number of AMY1 (Lazaridis et al. 2014, 409).
Placing this aDNA analysis in the perspective of NCT, it can be argued that this work contributed to our understanding of past human diets by suggesting that the culturally-induced change from a hunter-gatherer diet to a high-starch agrarian diet may have caused a positive selection pressure on the AMY1 copy number, further providing evidence as to how a culturally-induced dietary change can influence the evolution of our genome. However, further analysis is required from a variety of ancient specimens originating from both within and outside of the Neolithic to provide more solid evidence for the temporal relationship between an increased starch consumption and AMY1 copy numbers, and therefore engaging with the theory of NCT further.
Inchley et al. (2016). investigated the temporal element of amylase selection further, questioning if the selection occurred before the Neolithic. This study analysed a world-wide sample of 480 high coverage extant human genomes for the copy numbers of AMY1 (Inchley et al. 2016). Application of the 1000 Genome Project’s accessible genome strict mask were also used to retain information from regions that can be uniquely mapped using Illumina short reads (Inchley et al. 2014, 8). Published aDNA sequences of a Neanderthal (Prüfer et al. 2014) and one Denisovan sample (Meyer et al. 2014) were also used. Further analysis of the 50kbp genetic windows surrounding the AMY1 locus suggested a low polymorphism in samples when compared to ancestral sequences (Ref? see Figure 5). Due to the low genetic diversity of the regions surrounding the AMY1 cluster both within and outside of Africa, it was concluded that the selective sweep post-dated the split between Neanderthals, Denisovans, and humans, but originated before farming. The reason for the selective sweep was attributed to a dietary shift through new methods such as cooking (Inchley et al. 2014, 7). This work provides an example as to how aDNA analysis has contributed to our understanding of past human diets by providing a new insight on origins of the AMY1 selective pressure before the Neolithic. When placing this work in the context of NCT, it can be argued that this aDNA analysis suggests the positive selection for a high AMY1 copy number was not an inherited trait from archaic ancestors but occurred due to the culturally-induced change of cooking. Overall, this suggests a further way in which a culturally-induced change in relation to past diets has influenced selective pressures upon the human genome.
The work of Lazaridis et al. (2014) and Inchley et al. (2016) have provided some examples as to how aDNA analysis has contributed to our understanding of past diets. Each of these studies have provided a clear example as to how a culturally-induced change in the environment can cause a positive selection pressure upon the human genome. These works have also shown how this type of analysis can allow an understanding of the temporal origins of this selection, suggesting that whilst the advent of agriculture may have had some impact upon the evolution of a high copy number of AMY1, the original selection occurred with cooking.
3.2 Starch: aDNA potential
The potential of these studies could be increased by analysing the relationship between AMY1 and cooking at Neolithic cooking sites. Whilst the above studies have suggested that the high copy number of AMY1 originated with cookery, the impact of the Neolithic dietary change has not been fully understood. One potential site is Çatalhöyük, Turkey. Evidence for starch-based plant consumption is found at the site, including domesticates such as wheat, barley, and lentils (Carretero et al. 2017, 417). Evidence of cookery is also present, with starch in features related to cooking and storage (Carretero et al. 2017, 418). By testing the remains at the site for AMY1 copy numbers, the available aDNA data on AMY1 can be increased, therefore allowing a more detailed analysis on how this dietary change may have influenced the evolution of the human genome.
4. Dietary research: aDNA potential
An area for future research is dietary-related illnesses such as Celiac Disease (CD). CD is an intolerance to dietary proteins such as wheat, barley, and rye. CD is present in 1-2% of Western populations, and up to 5% in Saharawi populations (Zhernakova et al. 2010, 970). However, the disease has substantial morbidity and therefore would not be beneficial to human survival. In fact, this prevalence in modern societies is unusual as a gluten-free diet has only been available since the 1950’s (Zhernakova et al. 2010, 970). It is possible that aDNA analysis may provide an insight into why the disease is so prevalent in modern societies.
Zhernakova et al. (2010, 971) used modern datasets from GWAS of 8154 controls from four European populations combined with 195 individuals from North African populations. The work was conducted with the aim of understanding if CD-susceptibility loci show signs of recent positive selection, thereby attempting to understand why CD has not undergone negative selection in past societies. The results found that in Europeans, significant signs of selection were observed on the alleles rs17810546*G (IL12A locus), rs917997*A (IL18RAP locus), and rs3184504*A (SH2B3 locus); all of which are CD-associated (Zhernakova et al. 2010, 972). Meanwhile, in Saharawis, positive selection was observed for rs3184505*A (SH2B3) locus and rs917997*A (IL18RAP locus) (Zhernakova et al. 2010, 972). The study interpreted the results to suggest that the CD-associated loci of IL18RAP, IL12A, and SH2B3 may have undergone positive selection (Zhernakova et al. 2010, 970). This is because IL12A is involved in cytokine responses to protect against infections, whilst SH2B3 plays a role in an innate immune response (Zhernakova et al. 2010, 976). Overall, this study suggests that the dietary-associated disease of CD has not been selected against due to the genetic hitchhiking of alleles associated with immunological response. Furthermore, Mathieson (2015, 501) has alluded to the prominence of CD being a result of genetic hitchhiking to protect against ergothioneine deficiency upon a change to a Neolithic diet . These studies have shown that the dietary-related disease of CD may be prevalent in modern societies due to the positive impact of genetic hitchhiking. By understanding these beneficial alleles in a modern sense, it will be possible to trace the evolution of CD-associated hitchhiking in an ancient context.
One of the ways in which this type of analysis could be extended is by using population-scale and clinical databases such as genomAD to further understand the frequency of these linked loci in the extant population. Another source of published data are Medical datasets such as BioBank Japan and UK Biobank, which map the genetic associations of complex traits and diseases. Using published datasets to further understand CD-associated loci and alleles, it may be possible to earmark these variants to be studied in an ancient genetic context. By allowing the identification of any alleles which have undergone positive selection in a modern context, it may be possible to investigate these beneficial alleles in an ancient dietary context, and therefore understand how genetic hitchhiking can outweigh the impact of past dietary changes.
5 In conclusion
Through the defining of specific case studies, this review has shown how aDNA analysis has contributed to our understanding of past diets through the study of lactase persistence and starch consumption. The greatest contribution was an increased appreciation of the relationship between the genome and culturally-induced dietary changes, as explained through NCT. The case studies were used as a platform to suggest how this type of analysis could be utilised, nominating potential key sites for further analysis. A call for further research into dietary-related diseases was also suggested with the belief that further research may aid understanding of the ancient origins of these diseases, and therefore the relationship between human dietary-related diseases and genetic hitchhiking.