Unlocking the Past Page 6
Their discussion put forward a number of ways of accounting for this unusual result. It may have been that a stray dinosaur had found its way into Pääbo’s lab and contaminated the equipment in some way. A second possibility was that at some point in our ancestry hybridization with a dinosaur was involved. I am sorry to say that Pääbo’s group favoured an explanation far less evocative than either of these. This third possibility was that human DNA carrying the confusing ‘fossil’ genes had contaminated a few of Woodward’s samples. The various comments in the May edition of Science all pointed in the same direction. However meticulous Woodward’s experiments, they had not been immune to the contamination problems to which ancient DNA analyses were particularly susceptible. The DNA from dinosaur bones took as swift a plunge in credibility as had the Magnolia leaves. As with those fascinating Clarkia beds, the plunge left a lot of questions unanswered. Scott Woodward’s bones did seem interesting. What surviving biomolecules were contributing to the waxy texture? What were those surviving cellular structures, and what molecules within them were picking up the microscopy stains?
We now know enough about the chemical fate of bones to show that it is unlikely DNA will remain accessible to our current PCR methods in bones as old as these. The collagen matrix of the bone would long since have turned to soluble gelatine, leached away and opened up the bone’s interior to water and air. Only at the very lowest temperatures could bones have escaped this destiny, and bones more than 2 million years old are not going to be found in permafrost. If the waxiness of Scott Woodward’s Cretaceous bones indicated some kind of lipid accumulation, then some adjustments to this rule may be necessary. However, there is no evidence of this, especially now that it seems that none of his forty-two samples yielded anything other than contaminant DNA. More of a question mark hangs over the mineral fraction of the bone, the impure crystals of calcium phosphate or apatite. Matthew Collins, who has made a close study of the chemistry of bone, has speculated that the mineral surface itself provides the most durable depository for long-chain molecules such as DNA. If they survive here, as some are convinced they do, then they survive because they are very tightly bound. Here lies the problem. The same tight protection that may prevent chemists and biologists from seeking them out in the normal course of events also prevents analytic scientists from separating and identifying them. In the final chapter, we shall return to some recent attempts to overcome this problem. In the meantime, however, the time frame for ancient DNA analysis stretches back not much further than the Siberian woolly mammoth, 50,000–perhaps as much as 100,000–years.
This time span takes us back as far as the final stages of the human evolutionary drama, and the disappearance of the last hominids other than our own species. It extends beyond the appearance of what prehistorians call ‘modern behaviour’, vividly expressed in Stone Age rock art. It encompasses a period in which humans changed the genetic configuration of several other species, fuelling their most significant transformation of the ecosystem, through agriculture. Within a time scale that is shorter than the dinosaur hunters had hoped for, but which is long in terms of our own species’ history, ancient DNA can reach widely into the archaeological record. In just one lifetime, both archaeology and the life sciences have hurtled forward. Within living memory one discipline was all pots and stones, the other long lists of Latin names. Now they both converged on the code embodied in life’s most remarkable molecule, and the information flowing from variations in a sequence of four chemical bases. Like the range of text that can flow from a mere 26 letters in the alphabet, that information seemed boundless.
afterword
Several of those who were early career researchers at the time of the Ancient Biomolecule Initiative have gone on to lead research groups in Britain and abroad. On our steering committee, that major pioneer of the field, Svante Pääbo went on to take a lead role in the establishment of the Max Planck Institute for Evolutionary Anthropology in Leipzig. It was here that some of the most dramatic developments in our understanding of the human family would take place, explored in the following afterword. The Institute is an exciting place, a modern airy space just outside the historic city of J. S. Bach and Felix Mendelssohn. Its spaces are designed both for primary scientific work, and as importantly, for intellectual interaction, with central areas in which biomolecular archaeologists, geneticists, palaeoanthropologists, and linguists can engage and compare notes. Although the field is now much better established, unpredictable conversations across disciplines remain key to its continued success. Right in the centre of all this activity is a surprising structure that at first looks out of place. Pääbo built in the initial design of the Leipzig lab a climbing wall that dominates the building’s central well. His mind may be focused on the next human genetic challenge, but a part of his heart remains in his beloved Alpine mountains.
Several of the next generation of researchers who probed the vast Northern Arc from Siberia to Canada have also moved on to august institutions in which to become prominent leaders of their own research fields, and have been at the heart of some major initiatives in biomolecular archaeology. Willerslev has for several years directed Copenhagen’s Center of Excellence in Geo-Genetics, and is now joining the Cambridge University group. Alan Cooper, currently directing the Australian Center for Ancient DNA at Adelaide, was previously Director of the Henry Wellcome Ancient Biomolecules Centre at Oxford University, a position subsequently held by Beth Shapiro, before co-establishing a Paleogenomics Lab in Santa Cruz. The way in which each of these researchers now works with ancient DNA fragments has also changed radically. A game changer in the latter part of the last century had been the polymerase chain reaction (PCR), in which a way was found to artificially mimic the process of DNA replication that takes place in living cells. That important breakthrough has now been eclipsed by a whole series of developments to do with ‘high throughput’ sequencing (sometimes referred to as ‘next generation’ sequencing).
The contrast between conventional sequencing (often referred to as ‘Sanger sequencing’), and high throughput sequencing has some superficial resonances with the contrast between a craftsman’s workshop and a factory production line. In the former, a great deal of attention is paid to particular details of the items of production, each of which is fashioned with care. In the latter, the process is devolved to a series of repetitive actions, each undertaken in considerable numbers of different components. The analogy falls down in that the core theme of the factory is the assembly of components into a whole, whereas the core theme of high throughput sequencing is the disassembly of a whole, or of large fragments, into small components.
There are a series of distinct technologies involved, often gathered together, and the number continues to grow. As an example, let us take the technology known as ‘Illumina’. The first step of this technology is one of fragmentation, converting a particular source DNA sample into quite short lengths, each of less than 200 base pairs. The resulting multitude of short fragments are then exposed to molecular ‘adapters’, one part that attaches to the DNA strand itself, another part adheres to a glass slide, which thus accumulates a number of DNA ‘dots’. The whole group is then subjected to PCR amplification generating multiple identical copies at each dot.
At this point, the copies are lit up. The slide is flooded with DNA polymerase, and ‘colour-coded’ fluorescently labelled nucleotides, designed to attach themselves, one at a time, to the next exposed base in the sequence and to reveal their chemical identity by the wavelength at which they fluoresce. Because the stepwise nature of such attachment can be controlled, a sequential photographic record of the wavelength, and hence the nucleotide, can be made. The resulting output is a very large number of quite short DNA base-pair sequences, theoretically corresponding to the entirety of the original sample. The entire ‘book of life’ is in there, but in miniscule pieces as if it had passed several times through a shredder.
Each of the different novel technologies produces so
mething comparable to this, a vast array of very short sequence information, the multiple fragments that make up the longer sequence, if only one could get them all in the right order. Fortunately, we now have ways of doing that.
The history of DNA science runs more or less parallel to the history of information technology. The previous afterword emphasized the significance of both these fields moving forward at a dramatic pace. By the time Watson and Crick were celebrating their epochal discovery, the US government had already been using ‘computers’ of a sort, first for military purposes, then for census data. These were massive assemblies of equipment, filling a large room, but with much less capability than a single modern handheld device. Researchers of my generation can remember visiting such ‘mainframes’ with a bundle of ‘punch cards’ to carry out some fairly elementary statistical procedures on our data. The subsequent growth of computer-assisted analysis in biology has been spectacular. Bio-informatics is now in the front line of research, and it is the reason that, not only can we do something with the multitude of fragments described above, we can actually reassemble whole genome sequences, as was accomplished with the Thistle Creek horse. The bio-informatician now has a central, cutting edge role in our research, working with computers whose power may be of the order of a trillion times that of the pioneer computer used by the US government in Watson and Crick’s day.
The high throughput revolution has also transformed our approach to the perennial issue of contamination. Early researchers in the field were deeply concerned with cleanliness, controls, isolation, removal of surfaces, and so on, and all the while the possibility of contaminant molecules remained. While clean, controlled experiments are still very much the norm, the novel methods are affected by contamination in an entirely new way. The contaminants may be there, but are clearly visible as distinct fragments, alongside the core sequence to which attention is directed. We have become adept at recognizing contaminant ancient DNA.
Let us return to the ancient horse bone, introduced in the preceding chapter’s afterword, which Eske Willerslev’s team had recovered from Thistle Creek in Yukon territory. Samples from this foot bone were subjected to a pair of next generation technologies to generate 12.2 billion DNA reads, each read on average less than eighty base pairs in length. If early ancient DNA analysis was like putting a few jigsaw pieces together, the task now was more like assembling the entire board from a pile of sawdust. The care taken at every stage of that meticulous exercise in recognition, error correction, sequence assembly, and mitigation of contamination is explained in the group’s 2013 letter to Nature. Enough may be inferred from the horse’s genome to reconsider the depth of the species’ evolutionary history, its relationship with other taxa, and even its demographic fluctuation over time.
There is another feature of the progress of ancient DNA analysis which is illustrated by work on the Thistle Creek horse. This is the aspiration to target the evidence ever more precisely, an aspiration that ultimately leads to analysis of a single molecule. In principle, one of the sequencing technologies applied to the horse bone was doing that. The Helicos System proceeded in this by creating and observing fluorescently labelled base pairs detached from a single DNA strand. However, what may yet prove the most powerful minute-sample method available to ancient biomolecule research depends on something called a ‘nanopore’.
A ‘nanometer’ is a billionth of a metre, and a nanopore is a hole whose diameter is that order of magnitude. Such holes can be created in something similar to a biological membrane, essentially a sandwich of proteins and lipids. They can also be created in a sheet of silicon. Two features of nanopores are of considerable relevance to DNA research. First, the pores are of a size that may allow just one long chain molecule at a time to pass through them. Second, whatever molecular segment is passing through at any moment affects the passage of electricity through the sheet or membrane as it crosses the nanopore. Here we have the basis of the ultimate small sample analysis in DNA research. Nanopore DNA sequencing has been under development since 1995, but its impact on ancient biomolecule research lies in the future. Many in the research community are keenly alert to the possibilities.
High throughput sequencing has increasingly become a core feature of ancient DNA research, not just cutting down the time limits of analysis, and changing our approaches to addressing contamination but also in moving from analysis of short DNA sequences to whole genome studies. We could now follow this in a number of taxa, but can take the example from that central quest of biomolecular archaeology, understanding our curious cousins the Neanderthals. Leading that quest has been Pääbo’s Leipzig group.
A cave called Vindija in North Croatia provided one of the Neanderthal bones that had allowed the mitochondrial sequence to be explored. The skeletal fragments from the cave now allowed, provided two groups to publish (in 2006) further milestone papers on the path to reconstructing a full Neanderthal genome. In the Leipzig lab, Richard Green used a distinct form of high throughput technology (called 454 sequencing) to analyse one million base pairs of Neanderthal DNA. James Noonan, working from California, taking a rather different approach, demonstrated that a cloning base methodology still had much to contribute. The greatly enhanced body of genetic information from the two contrasting methodologies was impressively convergent. Four years later, Pääbo’s group had moved on at a great pace and were able to publish an initial draft of the Neanderthal genome based on four billion base pairs reassembled from the skeletal fragments of a human species, which was extinct.
However, the word ‘extinct’ must carry a footnote. Not surprisingly, an entire genome has caused us to radically revisit inferences drawn from fragmentary sequences of mitochondrial DNA. Among these inferences are the reproductive separation between Neanderthals and ourselves, which is by no means as great as was once inferred. A certain component of Neanderthal DNA is not as ‘extinct’ as the rest, and is alive and well within the genomes of living people, constituting 1 to 4 percent of our own genetic sequence, alongside fragments in genetic input from other, hitherto unrecognized curious cousins. To these we shall return in the following afterword.
3
our curious cousins
the far reach of ancient DNA
Tomas Lindahl had suggested a figure of between 50,000 and 100,000 years as his prediction of how far back accessible and amplifiable DNA might persist within a relatively cold bone. So where does that take us in terms of the human past? It goes beyond those most evocative expressions of the human mind encountered in cave art: these images appeared in different parts of the world less than 40,000 years ago. It also reaches back past the remarkable diversification of toolkits that occurred around the same time. It precedes any evidence of musical instruments and the elaborate rites of passage indicated by the treatment of the dead. Many of the key features that in our minds separate the ‘cultural’ humans from their more ‘natural’ predecessors appear in the archaeological record less than 40,000 years ago. The oldest DNA reaches back beyond that, to reveal an earlier episode in the human past.
Although the attributes of human culture listed above are not encountered in these earlier deposits, bones and a less diverse range of stone tools can be found. Recognizably human skeletal remains can be found in various parts of the Old World. They have no particular physical traits that cause us to place them outside the range of modern human forms. As far as we can tell, they are members of our own species. At either end of the Old World, in Europe and in east Asia, something else crops up. Today there is no question about what is human and what is not. We are a single species, with no barriers to fertility within our own species, and an absolute barrier to fertility between all other primates and ourselves. The position was less clear 50,000 years ago. In south-east Asia, skulls can be found similar to our own, except with a flattened skull and strong brow-ridge that no living human shares. In west Asia and Europe, a prominent brow-ridge appears in another group of specimens. Such traits survive in south-east Asia
until at least 27,000 years ago, as new scientific dating methods have recently shown, at a site along the Solo River at Ngandong, Java. The other bearers of strong brow-ridges were still alive 20,000 years ago further west in southern Spain. Archaeologists and anthropologists have long debated who these curious cousins were, and how different they were from us. What was a meeting with one of these hominid relatives like? Did the experience have the separateness of encountering a chimp or orangutan, with a similar shape but clearly of another species? Alternatively, was the meeting something more intimate, perhaps with exchange of things or ideas? Might it even have led to one bearing the other’s children?