Header image

john hawks weblog

paleoanthropology, genetics and evolution

Photo Credit: Split gill mushrooms. TheBeigeKnight CC-BY-SA 3.0

Mushroom-munching poplar-popping Neandertals

Neandertals ate mushrooms. That’s the conclusion of new work examining the DNA remnants in ancient dental calculus. Can we believe it?

Laura Weyrich and colleagues (2017) describe their work on dental calculus samples from five Neandertal individuals. The specimens include two from El Sidrón, Spain, two from Spy, Belgium, and one from Grotta Breuil, Italy (which did not in the end produce results). They also examined a larger number of ancient modern human specimens, including two Neolithic skeletons from Sudan, a number of South African Late Stone Age and Pastoralist Period skeletons, and a fairly large sample (n=75) of nineteenth-century Germans.

The last few years have seen an enormous increase in our knowledge of Neandertal dietary breadth. As recently as 10 years ago, while there was growing evidence that some Neandertals were using small game, shellfish, and other coastal resources, the going belief was that the Neandertal diet might have been composed almost entirely of meat—as much as 95% meat by some estimates.

This is understandable in light of the near-invisibility of plant foods in the archaeological record. It wasn’t only that plants don’t have bones. Neandertals and earlier people never made vessels or specialized artifacts that some later human populations used to collect and prepare plant foods. If Neandertals relied upon cooked grains, for example, you might expect them to have used grindstones once in a while.

Today, with new technology and some clever archaeological detective work, plants are no longer so invisible. Some Neandertals did cook and eat grains. Others relied on the starchy underground storage organs of plants like water lilies. They toasted pine nuts and lentils and ate dates. Some of this revolution in understanding Neandertal diet has come from microscopic remains of plant starches and phytoliths within layers of calcified plaque on their teeth, called calculus. Some has come from microscopic or biochemical examination of sediments within archaeological sites.

Now, ancient DNA from dental calculus is joining the party. I’ve heard some experts are skeptical of these new results. So what do they say and what are the weaknesses that we should know about?

Neandertal oral microbiome

Probably the main aim of the study was to characterize the bacterial oral microbiome of the Neandertals. Calculus includes genetic material from the bacteria that live in the mouth, and setting aside contamination, this bacterial signal makes up the vast majority of the DNA reads from the ancient specimens. Weyrich and colleagues found that the bacterial communities in the mouths of Neandertals were most similar to the African hunter-gatherers that they sampled. The single wild chimpanzee that they sampled also had a similar oral microbiome by this assessment. All these “wild” food eaters show a very different bacterial composition than the humans in the sample, both ancient and modern, who eat agricultural diets.

The paper includes the whole-genome sequencing of a species of archaea within the calculus of the sample from El Sidrón 1. As this specimen lived around 48,000 years ago, they call it the “oldest microbial draft genome” yet assembled.

Date estimates using a strict molecular clock place the divergence between the M. oralis strains of Neanderthals and modern humans between 112–143 ka (95% highest posterior density interval; mean date of 126 ka) (Fig. 3b; see Supplementary Information). As this is long after the genomic divergence of Neanderthals and modern humans (450–750 ka), it appears that commensal microbial species were transferred between the two hosts during subsequent interactions, potentially in the Near East.

This is fascinating if true. The claim is that an African-derived oral commensal species colonized the mouths of Neandertals in the Near East sometime after 126,000 years ago, then was carried to the furthest reaches of western Europe through sparse Neandertal populations, ending up in a Spanish Neandertal some 48,000 years ago.

Earlier studies have pointed to evidence of possible exchanges of human pathogens or parasites among archaic human populations or species. Maybe the most notable is the evidence of louse transfer. Last year it was suggested that human papillomavirus had been exchanged from Neandertals into modern human populations (“Today’s genital warts came from trysts between Neanderthals and Homo sapiens). Weyrich and colleagues are the first to document a possible exchange of a microbial species using ancient DNA evidence.

Still, before accepting this conclusion, we need to have greater confidence in the divergence time of these microbial strains as estimated from their DNA. Assembly of more microbial genomes from the Neandertals might yield a lot more information about such interactions.

These are examples of the cool stuff in the paper. A metagenomic library has many dimensions of variation to explore, and the paper can only summarize many of them.

Neandertal food DNA

The one part that got a lot of news coverage was the idea of “vegetarian” Neandertals.

This part of the study reflects a very small fraction of the DNA recovered from the calculus, the reads that map to some eukaryotic (as opposed to bacterial or archaean) genome. To get an idea how small a dataset this is, the shotgun sequencing of the Spy 2 individual yielded more than 17 million sequence reads, after filtering out more than 80% as probable contaminants. Only 532 of these reads mapped to possible eukaryotic dietary sources.

But of those 532 reads, 62% mapped to sheep and 34% to rhinoceros. That’s a pretty suggestive result, particularly when combined with the archaeology of the site, which includes sheep and woolly rhinoceros bones.

Of course, the presence of these other bones in the site itself prompts a question. Could the evidence of DNA from these species actually reflect ancient environmental contamination from within the sediment of the site? That’s a very hard hypothesis to test, since the site was excavated more than a hundred years ago. Although Weyrich and colleagues took precautions against including DNA from the surface of the calculus samples, the study cannot rule out the possibility of DNA uptake from ancient sources.

The Spy 2 calculus sample also contained DNA that maps to the genome of the grey shag mushroom, an edible species.

Grey shag mushrooms. Andreas Geminder CC-BY-NC-SA 3.0.

Weyrich and colleagues found that the calculus of the Spy 1 Neandertal was dominated by contaminating modern DNA sequence, so conclusions about diet from eukaryotic sequence reads are really not warranted.

Calculus samples from both the El Sidrón individuals contained DNA that maps to pine trees. The El Sidrón 1 calculus sample had a greater diversity of eukaryotic DNA, including sequences mapping to the split gill mushroom, poplar tree, and a species of moss.

The presence of such species, plus the lack of non-human mammalian DNA in the El Sidrón samples, gave rise to the “vegetarian Neandertal” idea in the press. That’s a vast overstatement of the data, since any one of these individuals would have had a diet including scores of eukaryotic species, even though only a tiny number were recovered from their calculus. Calculus data cannot document dietary breadth or the average diet, it can only document a few of the species that were eaten or were otherwise processed in the mouth. We already know from previous microscopic analysis of Spy and El Sidrón calculus samples (Henry et al. 2011, Hardy et al. 2013) that both sets of Neandertals had consumed starchy plant foods, as evidenced by abundant starch granules. These starchy plants are not present in the DNA data. So the ancient DNA in calculus is providing additional data that does not necessarily duplicate what is visibly present as microfossils.

The moss is interesting. Laura Buck and Chris Stringer (2014) suggested that one way Neandertals might have gotten plant foods into their diet is by eating the stomach contents of their prey animals. This is a practice in some recent hunter-gatherer groups living in Arctic or extreme northern environments. Poplar bark might also have gotten into the diet in this way.

Weyrich and colleagues speculate that the poplar bark may instead be further evidence of the use of medicinal plants by Neandertals:

Our findings support previous suggestions that El Sidrón 1 may have been self-medicating a dental abscess. This was the only individual whose calculus included sequences corresponding to poplar, which contains the natural pain-killer salicylic acid (the active ingredient in aspirin), and also notably contained sequences of the natural antibiotic producing Penicillium from the moulded herbaceous material. The sample from this individual also included sequences matching the intracellular eukaryotic pathogen microsporidia (Enterocytozoon bieneusi), which causes acute diarrhoea in humans, indicating another health issue that potentially required self-medication.

Another possibility is that the Neandertals were making artifacts or picking their teeth with poplar wood. The authors attribute the presence of pine DNA to consumption of pine nuts; charred pine nuts have been found at other archaeological sites. It’s also conceivable that the Neandertals were orally processing artifacts that involved pine wood or tar. Radini and colleagues (2016) last year reported finding fibers and compounds within the calculus of El Sidrón Neandertals compatible with conifer wood, and suggested this was non-dietary processing of wood for utilitarian purposes. The DNA traces may be tracking the same activity.

Teeth were tools, too. Photo: Stanley Zimny CC-BY-NC 2.0

Reasons for caution

I’ve seen some skepticism expressed about this study in my social networks. It is important to consider the limitations of this kind of work.

The eukaryotic data represent only a very small fraction of the sequence reads within the filtered datasets. This raises the possibility that a small fraction of contaminating sequence might generate such results. These are not singular reads. There are 23 reads mapping to split gill mushroom within the El Sidrón 1 calculus sample, for example. But it will take some careful comparisons to see whether they might be explained by other sources than dietary mushrooms.

There are two particular issues that should lend caution to the dietary interpretations. One issue is represented well by the sequence hits for DNA from a tick species (Ixodes scapularis) in the living human calculus sample that was newly examined in this study. Weyrich and colleagues aver that this individual was probably not eating ticks. Instead, this tick species’ draft genome likely incorporates contaminating human DNA reads which have been wrongly assembled together with the tick’s own DNA. It’s also possible that the tick genome draft mistakenly includes microbial DNA from the tick’s own microbiome. Such incorporation of contaminating DNA into genome drafts is a surprisingly common problem in genomes that are available in databases today; there has been a lot of sloppy genome assembly, particularly for arthropod genomes. The authors filtered their data against known human sequence contaminants within such genomes, but they did not rule out all such effects (as evidenced by the tick).

The second issue that concerns me is that nobody has yet sequenced the genomes of many potential food species of Neanderthals (and recent human hunter-gatherers). A short read that maps to split-gill mushrooms may indeed be from that species, but it is possible that relatives of this species, even quite distantly related, might actually be the source of the DNA. They just haven’t been included in databases yet.

It’s not likely that DNA from a bison will map to a mushroom. But many unmapped reads within the calculus data may represent species that haven’t yet been studied at the whole-genome level. As more and more potential food species are added to genome datasets, they may yield more (or better) hits from these same sequence reads.

Setting aside these issues of sequence data quality, for me the biggest reason for caution is our lack of actualistic data on calculus uptake of dietary DNA. The study provides no indication of how DNA trapped in the calculus of living people reflects their diets. We don’t know whether some kinds of foods are more likely to be taken up than others, whether foods are representative of certain times (when calculus is more likely to calcify, for instance) or whether some food sources are more durable in their DNA preservation within calculus (for example, woody tissues). We do know that trace microfossils that are present in some calculus samples represent species with no DNA evidenced in other calculus samples from the same individuals. So there’s a lot of information loss here.

What I’d like to see is a lot more work on living people with varied diets. Track what they eat, and see what species show up in their calculus. Of course, that’s tricky within populations that tend to remove calculus from their teeth by dentistry. Every biological comparison between living people and ancient people has to face the many differences in lifestyles that have emerged over the years, and calculus formation has a lot of variation among populations.

Bottom line

Of course, all kinds of calculus research, including previous studies that have shown plant consumption and cooking in Neandertals (e.g., Henry et al. 2011), face the limitation that calculus does not sample dietary breadth. What remains in a single calculus sample is only a tiny subset of the dietary behavior of any species. A sample may demonstrate some of the dietary lifeways of past peoples, but it is helpful to have other sources of evidence as well.

When it comes to total diet composition, the best evidence we have right now is for the Spy Neandertals. Last year, Naito and colleagues (2016) looked at isotopic evidence for diet in the Spy Neandertals, using an approach that considers the stable isotope contribution of individual amino acids within proteins. This finer-scale consideration of the sources of dietary protein was consistent with up to 20% of the protein composition of these individuals’ diet coming from plants. As Naito and coauthors point out, plants are much lower in protein composition than meat, so the total dietary fraction made up by plant sources would likely have been higher. That puts the Spy individuals into the range of modern human hunter-gatherer peoples at lower latitudes.

Different Neandertal groups had different dietary compositions, as reflected by microwear evidence from the teeth (El Zaatari et al. 2011). That means that the El Sidrón Neandertals really may have had a higher intake of plants than the Spy Neandertals, although isotopic results that would point in that direction are not yet available. I wouldn’t expect that any of them were vegetarians. The El Sidrón skeletal remains are thought themselves to be the result of cannibalism, and recent work on the El Sidrón microwear suggests a mixed diet (Estalrrich et al. 2017).

But what’s clear is that the overall plant use by Neandertals is much greater than anthropologists believed 10 years ago.


Buck, L. T., & Stringer, C. B. (2014). Having the stomach for it: a contribution to Neanderthal diets?. Quaternary Science Reviews, 96, 161-167.

El Zaatari, S., Grine, F. E., Ungar, P. S., & Hublin, J. J. (2011). Ecogeographic variation in Neandertal dietary habits: evidence from occlusal molar microwear texture analysis. Journal of Human Evolution, 61(4), 411-424.

Estalrrich, A., El Zaatari, S., & Rosas, A. (2017). Dietary reconstruction of the El Sidrón Neandertal familial group (Spain) in the context of other Neandertal and modern hunter-gatherer groups. A molar microwear texture analysis. Journal of Human Evolution, 104, 13-22.

Hardy, K., Buckley, S., Collins, M. J., Estalrrich, A., Brothwell, D., Copeland, L., ... & Huguet, R. (2012). Neanderthal medics? Evidence for food, cooking, and medicinal plants entrapped in dental calculus. Naturwissenschaften, 99(8), 617-626.

Henry, A. G., Brooks, A. S., & Piperno, D. R. (2011). Microfossils in calculus demonstrate consumption of plants and cooked foods in Neanderthal diets (Shanidar III, Iraq; Spy I and II, Belgium). Proceedings of the National Academy of Sciences, 108(2), 486-491.

Naito, Y. I., Chikaraishi, Y., Drucker, D. G., Ohkouchi, N., Semal, P., Wißing, C., & Bocherens, H. (2016). Ecological niche of Neanderthals from Spy Cave revealed by nitrogen isotopes of individual amino acids in collagen. Journal of human evolution, 93, 82-90. doi:10.1016/j.jhevol.2016.01.009

Radini, A., Buckley, S., Rosas, A., Estalrrich, A., de la Rasilla, M., & Hardy, K. (2016). Neanderthals, trees and dental calculus: new evidence from El Sidrón. Antiquity, 90(350), 290-301. doi:10.15184/aqy.2016.21

Weyrich, L. S., Duchene, S., Soubrier, J., Arriola, L., Llamas, B., Breen, J., ... & Farrell, M. (2017). Neanderthal behaviour, diet, and disease inferred from ancient DNA in dental calculus. Nature. doi:10.1038/nature21674