Header image

john hawks weblog

paleoanthropology, genetics and evolution

Photo Credit: Rainy day at Amud Cave, Israel. John Hawks CC-BY-NC-ND

My essay in The Scientist covers recent natural selection in humans

The Scientist has published a current essay I wrote about the evidence for recent and ongoing natural selection in humans: “Humans Never Stopped Evolving”.

Natural selection is tricky to catch in action. As Darwin put it, “A grain in the balance will determine which individual shall live and which shall die.” The grain in the balance—the slightly increased chance that organisms carrying one gene variant will fail in the struggle for existence—is the cost of selection. It is almost invisible, only becoming statistically evident when viewed across thousands of individuals, who may display only subtle differences in the affected character.
In the human population, the toll of natural selection is hidden within millions of deaths and births around the world every year. Everyone dies, many tragically young. And while obvious patterns sometimes emerge from early deaths—certain diseases, traffic accidents, drug overdoses—these are often challenging to connect to the action of genes. Likewise, only by comparing the genes of parents with those of childless people, and the genes of large families with those of small families, can we begin to understand how natural selection is acting on births.

I present a few of the classic examples, including lactase persistence and the Duffy blood group. Then I turn to consider the evidence of selection on continuous traits, much of which has just been emerging during the past couple of years. This is an exciting area right now, and there is much more to say about it.

I also struck a deal to make this article available via a Creative Commons license, and I will be reposting it after a couple of weeks.

Earnest Hooton, discussing the Taung Child in Up from the Ape (p. 284):

More temperate in our Australopithecine enthusiasm than Dart, we nevertheless must admit that here was a near-man (although still predominantly an ape) that failed of his destiny, if that was ultimately human. On the basis of teeth alone, it might have been claimed that Australopithecus was human. However, we must now admit that teeth do not make the man, although I am inclined to think that the loss of them may ultimately unmake him.

New Scientist this week has published a short opinion piece that I wrote about ongoing human evolution: “Are humans still evolving? Growing evidence suggests we are”. (unfortunately requires login) I discuss a recent research paper by Jonathan Beauchamp that found a negative selection gradient with increasing educational attainment in American adults:

Beauchamp tapped into the 20,000-person US Health and Retirement Study, which is augmented by genetic information from subjects. Looking at people born between 1931 and 1953, he found that in both men and women, higher educational attainment was correlated with having fewer children.
That much may seem obvious. For a century, Americans have been foregoing family size and early child-rearing for more education. Globally, this change is part of what is known as the demographic transition.

I also discuss some of the limitations of this study.

We will be seeing more and more results that show selection in these large-scale health studies that are incorporating genetics.

Probing the 'seven biggest problems facing science'

Vox writers Julia Belluz, Brad Plumer, and Brian Resnick compile the results of a survey they sent to scientists around the world, with the single question: “If you could change one thing about how science works today, what would it be and why?” Their resulting article points to several issues of common concern: “The 7 biggest problems facing science, according to 270 scientists”.

Many of these are issues that I’ve discussed here on the weblog over the years. Paleoanthropology of course has its own distinctive spin on several of the big challenges facing scientists worldwide, but in general there is a lot of overlap.

What I find interesting about the list is thinking about the ways that global challenges for science may be overlooked in the study of human origins. For example, the second problem discussed by Belluz, Plumer and Resnick is “Too many studies are poorly designed. Blame bad incentives.” They’re referring directly to issues like p-hacking and fishing for positive results, exacerbated by the use of too-small samples with weak statistical power.

Our respondents suggested that the two key ways to encourage stronger study design — and discourage positive results chasing — would involve rethinking the rewards system and building more transparency into the research process.
"I would make rewards based on the rigor of the research methods, rather than the outcome of the research," writes Simine Vazire, a journal editor and a social psychology professor at UC Davis. "Grants, publications, jobs, awards, and even media coverage should be based more on how good the study design and methods were, rather than whether the result was significant or surprising."
Likewise, Cambridge mathematician Tim Gowers argues that researchers should get recognition for advancing science broadly through informal idea sharing — rather than only getting credit for what they publish.

For many questions, paleoanthropology is the very definition of small samples with weak statistical power. This has its downside. Journals sometimes publish results on fossil hominins that would be meaningless in most fields, simply because everyone acknowledges the difficulty of doing better.

Many paleoanthropologists have seriously grappled with the problem of small samples. We do not shy away from replication studies and negative results. A negative result to a well-formulated research question can be just as interesting in the context of human evolution as a positive one. Papers that report a failure to replicate earlier results, or that emphasize how methods fail to support conclusions, are a regular feature of paleoanthropology research. Robust published debates are very common. These are all areas in which the study of human evolution is ahead of most other fields of science.

However, even though paleoanthropology may be better than many other fields when it comes to replication, we face the problem that some measurements cannot be replicated by independent scientists because the data are inaccessible. This is where, as a field, we fail to meet the standards that are common in other areas of science.

Joanna Klein in the New York Times reports on a new study by Lori D’Ortenzio and colleagues examining the lifetime history of rickets in “Old Teeth Tell New Stories About People Who Didn’t Get Enough Sun”

They found that a 24-year-old man who had been buried in Quebec between 1771 and 1860 had suffered four bouts of rickets in his short life: twice before the age of 2, once again around the age of 6, and again, in a somewhat severe episode, around the age of 12. Evidence of this final episode in his third molar correlated with an abnormal curvature in his tailbone that only could have developed around the same time.
“We were able to see inside that tooth, what was housed in there, years ago,” said Dr. Brickley.
They reached this precision because teeth develop at different rates and leave behind concentric circles like tree rings over time. The researchers could look at the abnormalities within those layers to estimate occurrence and severity.

The paper applies a really cool approach that relies upon the failure of dentin to mineralize properly during episodes of rickets. That makes it possible to look for distinct times during the childhood of individuals when they were suffering extreme vitamin D deficiency.

Today, although many individuals do not get the level of vitamin D that would lead to optimal health and development, relatively few suffer altered bone growth to the extent diagnosed as rickets. In skeletal samples of past populations, there are some very extreme cases of rickets. This new approach may help develop a better appreciation of how rickets manifested in past populations, and what the actual challenges were for vitamin D absorption during development.

A new Kickstarter-funded documentary is coming out about Dr. John Brinkley, the Kansas quack best known for his goat gland operations during the Roaring Twenties and into the Great Depression. The film is Nuts!. It seems to take an unusual approach, at first going into the story with a sympathetic view to the doctor, through animated reenactments, then exposing him as a huckster.

Andrew Lapin of National Geographic has an interview of the film’s director, Penny Lane: “The Bizarre History of a Bogus Doctor Who Prescribed Goat Gonads”. Of course, there is the question that everyone asks: What exactly was the operation?

So he just kind of stuck it in there under the skin, and if you were the patient, you could feel there was something in there. And if the surgery worked at all, that was the secret to it: that you knew you’d had it, and you were told that it worked. It was the power of suggestion.

It’s not really very far from things we see advertised today. Just tonight we were listening to the radio and heard an ad for a product: “It won’t bring you back to what you used to be, it will make you more of a man than you ever were!”

This is a part of history I love, being a native Kansan myself. As the American Medical Association tried to put an end to his goat gland operations, he made a serious run at the Kansas governorship, starting a media empire in radio in the progress. Then, when the government shut down his nationwide radio station, he moved it across the border, starting the phenomenon of “Mexican radio”, super high-powered stations that could beam their signal across the continent.

The 2009 book that inspired the film was Charlatan: America’s Most Dangerous Huckster, the Man Who Pursued Him, and the Age of Flimflam, by Pope Brock. I really enjoyed this book, and if you want a very readable account of this history, I strongly recommend it.

Managing science amid the complexity of specialization

Larry Moran comments on a recent study that attempted to quantify the increasing complexity of scientific papers: “The scientific literature is becoming more complex”. The meat of the study is that the average scientific paper now has more data, tables and figures—almost double the number twenty years ago.

The number of authors on the average paper has also increased over time, and is today correlated with the “impact factor” of the journal. This suggests a pressure to increase complexity by increasing specialization and drawing together more specialists in the preparation of high impact research.

Larry comments:

I think the major consequence is the lack of responsibility of individual authors in a multi-author study. With increased specialization, there are fewer and fewer authors who see the big picture and who are capable of integrating the results from several subspecialties. The fact that the studies include work from several highly specialized techniques that only a few people understand also makes it harder for the average reader to evaluate the paper.
It's likely, in my opinion, that many of the authors on the paper don't fully understand the techniques being used by their colleagues. This is a big change from the science I grew up with.

It definitely is a big change from the science we grew up with.

We have moved beyond the world of multi-author research into the world of pluri-author research. Sure, we’ve seen papers with fifty or a hundred authors (or even vastly more) being published for a long time. But now they are the mode for a certain kind of important work.

I’ve been fortunate to have been involved in several effective large-scale collaborations. This includes the most exciting work I’ve done. The right combination of people will think of things that no single person or laboratory would have done, resulting in better and broader work. They can take on “big picture” questions that are beyond the scope any single investigator or laboratory could achieve.

I’m not alone, I know many anthropologists and geneticists who manage the big picture by communicating effectively within a large team of specialists. We have not reached the limit of what we can do integrating results from different areas of research.

Still, this complexity is one of the great challenges of science today. It is not always successful. In pluri-author papers, it can be a real challenge to keep track of what everyone has done.

More and more, there are published papers where the different methods used actually contradict each others’ results. Contradictory results are expected on a certain scale of work, but in those cases, the work needs someone with the right experience and training to put the contradictions into context.

Such generalist training is harder and harder to come by.

Being capable of managing and interacting effectively in such large-scale collaborations is not an innate talent, it is a skill that scientists can learn. Personally, I train my graduate students to be generalists, capable of conceiving and designing research that combines multiple specialties. I think a narrow specialization in the long run will cause students to miss out on great opportunities.

However, I recognize that many students are trained in techniques that can be very productive in their first few years, giving them a quick burst of publications and sometimes grant funding. This kind of scientist faces limits in the long run, as it is extremely hard to keep up in fields where research techniques may be rapidly changing. But many scientists maintain large successful careers as specialists, building work together with other scientists.

Looking at urban archaeology of the 'lost city' of Angkor

Annalee Newitz has a great article in Ars Technica that reviews a bunch of recnet archaeological work on the urban complex of Angkor Wat: “How archaeologists found the lost medieval megacity of Angkor”. This is one of several cases in which archaeologists are using large-scale LIDAR survey to learn more about the development and later abandonment of cities.

Archaeological researcher Piphal Heng, who studies Cambodian settlement history, told Ars that the LiDAR maps peeled back the forest canopy to reveal meticulous grids of highways and low-density neighborhoods of thousands of houses and pools of water. There was "a complex urban grid system that extended outside the walls of Angkor Thom and other large temple complexes such as Angkor Wat, Preah Khan, and Ta Prohm," he said. With the new data, scientists had solid evidence that the city of Angkor sprawled over an area of at least 40 to 50 square km. It was home to almost a million people. The scattered, moated complexes like Angkor Wat and Angkor Thom were merely the most enduring features of what we now know was the biggest city on Earth during the 12th and 13th centuries.

The famous, densely occupied city centers of such places, with their large temples and other civic structures, have obviously occupied a lot of attention from archaeologists in the past. However, focusing on these obvious structures gives a misleading picture of how the urban centers were maintained by the much larger population living at lower density over the surrounding area.

These good surveys from LIDAR and other data sources enable archaeologists to understand what they are sampling when they look at surrounding areas. They can focus excavation work on particular neighborhoods within the broader urban area, finding out whether they are diverse and understanding their economic basis:

The picture that's emerging of Angkor is much like a modern low-density city with mixed use residential and farm areas. As Evans put it to Ars, "in the densely inhabited downtown core there are no fields, but that nice, formally planned city center gradually gives way to an extended agro-urban hinterland where neighborhoods are intermingled with rice-growing areas, and there is no clear distinction between what is 'urban' or 'rural'." The city was a miracle of geoengineering with every acre transformed by human hands, whether for agriculture or architecture.

It’s pretty cool to find evidence of engineered landscapes extending out far beyond the massive structures. Geoengineering was a much greater accomplishment in many ancient societies than any of the large temple or funerary complexes that they created.

Traci Watson from National Geographic reports on a study of behavior in whales looking specifically at how members of social groups react to the death of an individual in their group: “Whales Mourn Their Dead, Just Like Us”.

In one case, short-finned pilot whales in the North Atlantic Ocean made a protective circle around an adult and dead calf. In another case, a spinner dolphin in the Red Sea pushed a young animal’s body toward a boat. When the vessel’s occupants lifted the carcass on board, the entire group of dolphins nearby circled the boat and swam off.

Many of the behaviors that they describe are similar to those seen in primates and other terrestrial social mammals like elephants.

Why the new Wellcome Trust open review journal may be the science publishing future

John Bohannan in Science writes one of many stories about the Wellcome Trust establishing a new open access journal, in which peer review follows the posting of preprints: “U.K. research charity will self-publish results from its grantees”.

Normally, peer review is anonymous and happens before publication of a paper. The charity’s journal, called Wellcome Open Research, will encourage researchers to post their work immediately on the site, as a full research paper or even just a data set. Only then does the peer review begin, and in this case the reviewers selected by the journal’s editors will be publicly known. "The transparent peer review process will encourage constructive feedback from experts," the Wellcome Trust's press release reads, "[focusing] on helping the authors improve their work rather than on making an editorial decision to accept or reject an article." Some other scientific journals and preprint servers such as arXiv similarly use postpublication peer review, although the concept has so far failed to be fully embraced by the research community.

It makes no sense whatsoever for funders to support scientific work, and then tolerate that work being behind a paywall that not even the funders can access, much less the public. Bohannan directs some attention to this aspect:

"This really is a potential game changer for a major funder to be taking control of the research output," says Paul Ginsparg, the Cornell University physicist who founded arXiv.org, the massive online scientific preprint server. He hopes that U.S. funding agencies will follow suit. "It would be a miracle."
From the point of view of Wellcome and other nonprofit groups that fund science, academic journals can be an expensive drain on time and money. Publication can take months or even years before anyone gets to read the output of the research they back, and with traditional subscription journals the reader then pays for the privilege.

This seems like the obvious move for large organizations funding science. If you want to increase the value of the work you fund, then you should make it easier to disseminate the results of that work to a broader number of researchers and the public. The preprint process is accelerating research by allowing others to build on the research faster.

Open review of preprints also has the potential to broaden the scope of outside review, bringing more voices into the process. But so far in biological sciences, it has failed to achieve its potential. Scientists may be reading and using results from preprints, but a robust commentary around new research has yet to emerge in these fields. Physicists and economists have long relied on pre-review dissemination of research, but it takes a cultural change to recognize these efforts.

Wellcome’s initiative may help shift the ground by taking some researcher fear out of the process.

The earliest scientific journals had editors who selected or solicited articles from experts, on the basis of interest to scientific readers. The experts of the time, in the eighteenth and nineteenth centuries, were gentleman philosophers, with science as an avocation, not a source of practical income. The best of the early editors were community builders, drumming up interest in scientific work at the same time they tried to advance the standards of scientific practice. This role of curating work continues today. Journals like Nature and Science, and even field-specific journals like the Journal of Human Evolution, have editors who select articles that they think will be valuable for their readers, both for community-building and for displaying scientific importance.

One of the big issues in science today is the fact that researchers are gaming their research outputs toward results that fit the values of journal editors. Editors and reviewers exhibit bias toward results that are “positive” in the sense of demonstrating statistical evidence for a hypothesis (as opposed to failure to disprove a null). They also like “counter-intuitive” results, because they seem novel as opposed to simply replicating what is already known.

Many people have pointed out that exactly these kinds of results are least likely to be true. Scientists are on a massive wild goose chase for false positives. “High-impact” journals print such results disproportionately often, and they pay little price when these results are later shown to be less than initially advertised.

The current system seems to work very well for certain journals, but poorly for funders who want to support solid research. Such problems have well-known solutions. Increase sample sizes. Pre-register studies, so that that researchers cannot change them based on what appears to be “significant” in a small sample. Require more from “counter-intuitive” results. So far, most journals have done very little to encourage these solutions.

A preprint server by definition does not select articles. With the arXiv and bioRxiv preprint servers, each submission undergoes a few checks at the time of submission to ensure that it fits the posting criteria, but those checks do not select articles on the basis of interest or perceived importance. Any scientific research that fits the preprint server’s remit will be accepted.

With Wellcome Open Research, the remit is any research funded by Wellcome. This creates a very interesting situation that we haven’t seen before: The curation of research for this venue occurs with funding decisions and not after research is completed.

This open approach is a funder-level equivalent of pre-registration. In essence they are saying: “We will fund projects and provide the infrastructure to publish them, irrespective of how they turn out.” They are making their own version of “impact” by disseminating both the work of researchers that they fund, and the review of that work. By doing so, they show the quality of science they are funding.

The criticism will be that Wellcome is making a walled garden. Sure, the research may look good on the surface, but it hasn’t faced the true competition of work in the major journals. Wellcome can do much to end this line of criticism. The benefit of open review is that its quality is on public display. But that means Wellcome needs to have excellent reviewers take part in the system. This is going to take a change in culture. The best reviewers have many demands on their time. Open review can allow referees to build a reputation for good community interaction, and what Wellcome can do is magnify those incentives. As a funder, Wellcome can demand something from its grantees that would benefit the community. But it should also think about carrots that will bring more reviewers into their system and create more engagement for funded researchers. A great community will build great science.

Will it work?

Obviously the scheme leaves a lot of wiggle-room for researchers to adjust their projects as circumstances require. The individual researcher still has many strong incentives to goose results to attain higher impact, and so far nothing is to stop them from submitting their favorite work to Cell.

Still, there are many advantages. I have found the eLife collaborative review model to be better than traditional review in many ways, and if Wellcome captures the benefits of a more open review process, most researchers will be eager to publish their work this way. Researchers and institutions worry that publishing in non-traditional venues will hurt their chances of obtaining funding. As a major funder, Wellcome can officially end that worry for their grantees.

I think we will see other funders following this lead, particularly those with strong public missions. It works against the interests of public-facing organizations for the work that they support to be behind a paywall, and at the same time, it is a huge suck of value to pay Elsevier or NPG to release work under a Creative Commons license. Organizations can add immense value by taking on the review and dissemination role directly.

Vegas Seven has a great interview with biological anthropologist Debra Martin: “Seven Questions With Biological Anthropologist Debra Martin”, touching on her work in bioarchaeology and how it relates to the public perception of forensics. This snippet of ongoing work caught my attention:

We’re interested in non-lethal violence [in the ancient Southwest]. We’ve been collecting data from skeletal [remains] from all over the ancient Southwest, trying to get a feel for differences across adult ages, differences between [genders] to see if there’s a pattern of non-lethal violence. When people migrated to other places, that’s where we start to see some of these head wounds show up. It’s really hard for migrants to enter into new locations. Migrants are highly stressed populations and are the subjects of more violence today, as we’re seeing in Syria.

Virginia Morell’s excellent biography of the Leakey family, Ancestral Passions, includes a great discussion of the aftermath of the innovation of potassium-argon dating which demonstrated that the Zinjanthropus skull was much older than anyone had assumed. This quote from Yves Coppens is well worth sharing, as he recounts the field’s usual reaction to any new discovery (snarky disbelief), and their subsequent actions (p. 196):

Everyone said the dates were good for the lava, but not for the skull. But as paleontologists are very clever as you know, they thought there must be many australopithecines in East Africa. So there was a sort of "bone rush" like the gold rush in the far West. Officially, the paleontologists were saying, "No, no, it's too old," but behind their backs they were packing their suitcases to go to East Africa. They came like locusts.

In 1930, Robert Broom commented on the age of the Taung specimen. This is one of the earliest instances I have found of someone claiming that a fossil is “too recent” to be an ancestor:

The little fossil ape skull that was found at Taungs five years ago is, in the opinion of many, the most important fossil ever discovered. It is manifestly the remains of an anthropoid ape somewhat allied to the chimpanzee, and of about the same size. But it differs from both it and the gorilla in a large number of characters, and in almost all these characters it resembles man. It thus seems highly probable that it is very near to the ape from which man sprang, and possibly a representative of the very genus. One difficulty has been our ignorance of the age of the cave deposit. The bone breccia found in most caves has proved to be of Pleistocene age, and if the Taungs cave is also of this period, then Australopithecus would be too recent to be a possible human ancestor, as man is known to have existed in the Pliocene.

From today’s perspective, the Taung specimen is believed to be a bit older than 2.5 million years old, but may easily be older or younger; Jeffrey McKee revisited the fauna in a 1993 article, and debate about the exact position and depositional circumstances represented by the skull continue today.

Of course, with today’s definition of Pleistocene extending to 2.5 million years ago, a date of that age is perfectly consistent with Broom’s statement–and many anthropologists might agree that the skull is “too recent” to be a lineal ancestor of today’s humans, by the same argument. I’ll have more to say about the concept of ancestors and geological age soon.


Broom, R. 1930. The age of Australopithecus. Nature 125:814. doi:10.1038/125814a0

McKee, J. K. 1993. Faunal dating of the Taung hominid fossil deposit. Journal of Human Evolution 25(5): 363-376. doi:10.1006/jhev.1993.1055

Nice piece from Kate Crawford in the New York Times about how predictive technologies used by Google and others go wrong when applied outside the context they were trained on: “Artificial Intelligence’s White Guy Problem”

If we look at how systems can be discriminatory now, we will be much better placed to design fairer artificial intelligence. But that requires far more accountability from the tech community. Governments and public institutions can do their part as well: As they invest in predictive technologies, they need to commit to fairness and due process.

The Pacific Standard is running a nice interview with anthropologist Barbara King, by Francie Diep: “How Do Gorillas Grieve?”. It touches on the recent killing of the gorilla, Harambe, at the Cincinnati Zoo, but broadens to consider the emotional lives of gorillas.

Q: What you’re saying reminds me that humans also evolved things like maternal love and aggressive responses, but that doesn’t mean people aren’t thinking, obviously.</strong>
A: That’s exactly right. I do not think it’s anthropomorphic to talk about grief in gorillas. Anthropomorphism is a projection of human qualities onto apes. Why would we say that grief is a human thing? I don’t think it is. I think it’s an animal thing, at least for some animals.