The Shadow King (33 page)

What concerns the critics most of all, though, is the type of DNA test that Zink and the team used. Ancient DNA researchers usually start a study by looking in a sample for mitochondrial DNA, which, if you remember, is much easier to find than nuclear DNA. Unusually, though, Zink’s team only gives results for the more challenging nuclear DNA. What’s more, rather than sequencing this DNA (which allows you to check, for example, that you’re not picking up DNA from more than one individual at once), the researchers analyzed it using DNA fingerprinting. This technique is useful for determining family relationships, but it is rarely used for ancient DNA studies because it is ultra-susceptible to confusion from contamination.

Microsatellites, the variable regions of the genome that DNA fingerprinting targets, consist of a short sequence—for example “GATC”—repeated several times. The number of repeats varies in different people. By using PCR to amplify particular microsatellite regions, then looking at the length of the resulting pieces of DNA, you can calculate how many repeats each microsatellite is made up of—the more repeats, the longer the DNA.

It’s a really cunning technique. The problem is that simply checking the size of PCR products offers no way to distinguish between ancient DNA and modern contamination. There’s no sequence to analyze, just a colored band or peak on a computer screen, which could just as easily come from a modern workman as an ancient pharaoh. To make matters worse, the PCR can slip on the sequence repeats, producing false “stutter bands” of varying sizes.

When working with modern, good-quality DNA, researchers can usually see which are the stutter bands and which are the real ones. But poor-quality samples of DNA are much trickier. For each microsatellite, the bands you’re interested in may or may not be present, depending on whether the DNA has survived well enough to amplify. They may or may not be mixed with DNA from one or more other individuals. And on top of that, you may see stutter bands. Teasing out which bands are real and which aren’t is fraught with difficulty. Get the answer wrong, and you’ll end up with a completely different fingerprint.

Even when used in criminal investigations—where it has helped to imprison or even execute plenty of people—DNA fingerprinting doesn’t necessarily give black-and-white results. Here, too, DNA from several people is often mixed together, and scientists are pushing the limits of the technique by chasing DNA profiles from smaller and smaller samples.

In an investigation carried out in 2010, for example, New Scientist magazine tested how objective the method really is.15 Reporter Linda Geddes took a sample of DNA evidence from a real crime scene—a gang rape in the State of Georgia—that had helped to convict a man named Kerry Robinson. She sent it, along with Robinson’s DNA profile, to seventeen analysts working in the same accredited U.S. government lab, and asked if they thought Robinson’s DNA was in the sample. Only one of the seventeen agreed with the original judgment that Robinson “could not be excluded,” while four said the evidence was inconclusive, and twelve ruled him out completely. As one attorney put it: “The difference between prison and freedom rests in the hands of the scientists assigned the case.”

For ancient DNA, where tiny samples of degraded, fragmented DNA are mixed with modern DNA that amplifies much more easily, the problem can be even worse.

IN OCTOBER 2010, a few months after the publication of Lorenzen and Willerslev’s letter, I call Carsten Pusch for his reaction. He’s clearly proud and excited about his role in the project, and is happy to talk me through the details.

As we chat, I check out Pusch’s website. Prominently displayed is a photo of him sitting next to Egypt’s antiquities chief Zahi Hawass at the press conference at which the DNA results were announced.* On the other side of Hawass are Ashraf Selim and Yehia Gad, the picture of respectability in ties and suits. The German geneticist makes a striking contrast against the three graying Egyptians, with long brown hair drawn back into a ponytail, a gold chain around his neck, and two small hoops through his left ear.

Pusch argues that the study owes much of its success to the way in which the royal mummies were embalmed. He believes that the very embalming materials that made the analysis of the royal mummies so difficult may actually have protected their DNA from degradation. “Nobody has thought about the components of the resin,” he says.

Members of the royal family would have enjoyed a particularly elaborate embalming process, he points out, perhaps explaining why their DNA is better preserved than that of other Egyptian mummies. As yet there is no direct evidence for this idea, but Pusch says he is working with chemists at the University of Tübingen to try to identify the materials involved, and understand how they might protect DNA.

Pusch’s colleague Zink, the slightly disheveled director of Bolzano’s mummy research center, is just as happy to discuss the project. He too argues that the mummification process, including the rapid desiccation with natron salt, probably helped to preserve the DNA beyond what would normally be expected. “The Egyptians really knew how to preserve a body,” he says.

Zink, a veteran of the field, was ready for the backlash against the study. In fact, “the skepticism we received was not as much as I had expected,” he says. But Pusch seems genuinely surprised and hurt by the reaction from critics such as Lorenzen and Barnes. “I don’t understand people’s harshness,” he says, after detailing the months of painstaking experimentation it took to coax DNA from the mummies’ bones. “These people have never worked with royal mummies. This is pioneering work.”

When I ask about contamination, both researchers say that they did pick up some contaminating DNA in their tests. They tested every microsatellite lots of times for each mummy, and got different results on different runs. According to Zink, they never got a complete profile with one run, so they built up each mummy’s fingerprint by looking at which bands came up most often across the different runs, using a “majority rule.” For example, when repeating a test thirty times, if they saw a particular band more than fifteen times, they’d judge that to be an authentic result.

Despite the messiness of the data, Zink and Pusch say they are convinced that the bands the team selected are authentic because they came up in different experimental runs, in biopsies from different sites on each mummy, and in two different labs.* They also point out that none of the female mummies tested positive for the male Y chromosome—suggesting they were not contaminated with DNA from male archaeologists—and that the mummies tested all had different (but related) profiles, so the results could not have come from the same source of contamination.

The concern is that the team must have been under huge pressure to find something big in those unpromising samples: from Discovery, who built a million-dollar lab in return for permission to film the research; from the domineering Hawass, whose reputation rested on the success of the project; and from the world’s waiting press. And in such circumstances, the use of DNA fingerprinting might make it all too easy for scientists trying to make sense of confused and ambiguous data to pick—in all good faith—the bands that made most sense for the family tree they were trying to construct.

The fact that all of the resulting DNA profiles made sense might therefore reveal more about the ability of the human brain to see patterns in random noise than it does about the family relationships of Tutankhamun and his kin. If so, that could explain, for example, why almost every anonymous mummy included in the study, even ones with no obvious archaeological link to Tutankhamun, ended up being identified as important members of his close family, and why several of Hawass’s hunches about the mummy’s relationships were found to be correct.

With so much room for interpretation, critics argue that the DNA tests should have been conducted “blind,” so that the researchers doing the analysis and working out the relationships between the mummies didn’t know which sample was which. They also want to see the raw data—those colored bands on the computer screen—to judge its quality for themselves (the JAMA paper only gives the final profiles that were eventually derived for each mummy). There doesn’t seem much chance of this, however. Zink says he’s happy to discuss the team’s methods and results with other researchers, but is reluctant to share the raw data, because the use of the majority rule means “there could be a lot of arguing.”

Something else the critics are keen to see is sequence data, particularly from the mummies’ mitochondrial DNA and Y chromosomes, as this would provide a valuable check on the family relationships revealed by the DNA fingerprinting. Zink and Pusch say that they do have some of this data, but chose not to include it in the JAMA paper as they are still working on it.* They insist that when they’re ready, they will publish this in a separate paper that investigates the genetic origin of the pharaohs.

“People are asking so many questions,” says Pusch. “We are able to do it. I just wish everyone would give us more time.”

Of course, none of the criticisms mean that the team’s results are definitely wrong. Gad, Zink, and Pusch appear to me to be talented and tenacious researchers working in very difficult circumstances. Maybe they really have pulled off the genetic feat of a lifetime. But unless they can provide the methodological details and raw data that will allow other experts to judge the validity of their work, there is, frustratingly, no way to know whether their results are real or not. A new history may have been written for TV audiences, but neither geneticists nor Egyptologists are ready to throw out their textbooks just yet.

There are some glimmers of hope for the future, however. Some recently published studies on animal mummies from Egypt hint that DNA does sometimes survive after all. The lack of convincing animal studies has been one reason why critics such as Gilbert are so skeptical that the DNA being claimed from Egyptian mummies is authentic. “It only ever ‘works’ on samples like humans that are really easy to contaminate,” he points out. But in a paper published in 2011, using rigorous methods and controls that even Gilbert approves of, a team amplified and sequenced short DNA sequences from a couple of mummified crocodile hatchlings, around two thousand years old.16 Another preliminary study reports using PCR to amplify DNA from mummified ibises.17

But the real breakthroughs may come from a series of new techniques collectively known as next-generation sequencing. These reduce reliance on the troublesome PCR amplification step, or even skip it altogether. Instead they sequence DNA in a sample directly, reading thousands of tiny fragments at once, then use a computer to stitch them all together into a longer sequence.

So far, next-generation sequencing has required access to sophisticated (and very expensive) equipment. But it is fast becoming more routine, and within a few years, probably all ancient DNA will be done this way. Because it avoids amplification by PCR, it is easier to estimate the amount of modern contamination in your sample, and you don’t have to worry so much about the contamination drowning out the ancient DNA you’re interested in. You sequence all of the DNA you’ve got in a sample, so you can see exactly what’s in there, including what species it is from, whether there is DNA from more than one individual or species, and whether it shows patterns of degradation that you’d expect from ancient DNA. Another big advantage is that you can look at much shorter fragments—down to about thirty base pairs long—which means being able to retrieve results from much older samples than before.

Using these techniques, the big labs such as Willerslev’s and Pääbo’s are looking again at ancient human samples, and they’re managing to read entire genomes for samples where this previously wouldn’t have been thought possible.

In 2010, Willerslev, Gilbert, Lorenzen, and colleagues used next-generation sequencing to read the full genome sequence of a four-thousand-year-old paleo-Eskimo from Greenland.18 Willerslev had spent two months digging in Greenland’s most remote northern tundra in 2006, looking for human remains, but with no success. Then, two years later, he found just the sample he was seeking—four tufts of hair from an extinct Eskimo group known as the Saqqaq, dug from the permafrost at Qeqertasussuk, on Greenland’s west coast. The whole time the hairs had been sitting in a plastic bag in the National Museum of Denmark, just minutes from his Copenhagen lab.

The resulting genome sequence showed that the closest living relatives of the Saqqaq are the Chukchis, who now live at the easternmost tip of Siberia. The researchers also deduced that the owner of the hair was a man with brown eyes, thick hair, and dry earwax, at risk of baldness later in life. It was the first whole genome of an ancient human (and at the time just the ninth human genome ever sequenced)—“an absolutely amazing piece of work,” according to Barnes.

It didn’t stand on its own for long, though. Within a few months, teams led by Pääbo published the genome sequences of a 38,000-year-old Neanderthal* recovered from Croatia19—proving once and for all that modern humans did interbreed with their Neanderthal cousins, leaving their trace in our genome today—and a girl belonging to a previously unknown human species of about the same age, whose tiny finger bone was found in a cave in the Altai mountains in Siberia.20

And in 2012, a team led by Zink published the genome of Ötzi the Iceman, showing that he had brown eyes, type-O blood, was lactose intolerant, and that his closest modern-day relatives now live in Sardinia and Corsica.21

So far, no one has successfully used these techniques on Egyptian mummies. But there’s every chance they could produce equally dramatic results. Going from 100-bp fragments down to 30 base pairs should make a huge difference to DNA survival time. Whereas 100-bp fragments are expected to survive only five hundred years in a hot Egyptian tomb, 30-bp fragments are predicted to last two thousand years or even longer.

“Everything becomes open again,” says Willerslev. As the techniques are still new, researchers don’t know yet exactly what will work and what won’t. Most of the genomes published so far have involved samples preserved in the cold, but Willerslev is now using next-generation techniques to extract DNA from various South American mummies, some of which have been preserved in warmer environments (as well as the hair of the famous Native American chief Sitting Bull, which he was presented by the family in a basement ceremony in Dakota, involving singers, drummers, and a medicine man). “Some [of the samples] are definitely working,” he says.