Ancient DNA from a recent descendent of a Neandertal-Homo sapiens coupling

Interesting article by Ann Gibbons in today's Science magazine. The article reports on a talk by Qiaomei Fu who sequenced ancient DNA from a young male mandible from a cave site in Romania. The mandible is said to be suspiciously robust, and it turned out to have a lot of Neanderthal nuclear DNA, including long segments that had not yet been disordered by cross-overs. The conclusion is that the individual was only a few generations away from a mating between a human and a Neanderthal. Both the date (42-37 kya) and the location (Europe) are curious for the interbreeding event. The scientific article is supposed to be in review. Can't wait!

Gibbons, Ann (2015). Ancient DNA Pinpoints Paleolithic Liaison in Europe. Science Vol. 348, Issue 6237, p. 847.

Hoyo Negro skeleton mtDNA

For those of you who haven't been keeping up on cave archaeology in Mesoamerica, a remarkably complete skeleton of late Pleistocene girl was discovered deep underwater in a cave near Tulum, Quintana Roo, a couple of years ago. Reports of skeletons and other artifacts in similar contexts have been floating around for years. Most of us presumed, I think, that these remains were quite early because the caves were flooded by post-Pleistocene sea-level rise. The cave near Tulum was named Hoyo Negro, and the remains of the young woman were nicknamed Naia. In 2014, her mtDNA sequence was reported along with dates suggesting the skeleton was 12,000 to 13,000 years old (1).

In today's issue of Science, Kay Prufer and Matthias Meyer of the Max Planck Institute for Evolutionary Anthropology argue in a comment (2) on the original article that mtDNA haplotype was incorrectly identified and was came from modern contamination because it did not show the level and type of damage they would have predicted.

The response (3) is robust. While conceding that the lack of damage to parts of the DNA sequence is perhaps unexpected, they argue that damage to ancient DNA is poorly understood and unpredictable. Therefore, it cannot be used by itself to prove or disprove that a particular sequence or segment of DNA is ancient. They also provide a reasonably compelling argument against contamination.

Both the comment and the response discussed whether and how DNA might be damaged  in different environments, including the tropical environment of Quintana Roo. What surprised me was the lack of any explicit consideration of the fact that the bones had been deep underwater for thousands of years. That seems like a noteworthy omission.

1. James C. Chatters, Douglas J. Kennett, Yemane Asmerom, Brian M. Kemp, Victor Polyak, Alberto Nava Blank, Patricia A. Beddows, Eduard Reinhardt, Joaquin Arroyo-Cabrales, Deborah A. Bolnick, Ripan S. Malhi, Brendan J. Culleton, Pilar Luna Erreguerena, Dominique Rissolo, Shanti Morell-Hart, and Thomas W. Stafford Jr. (2014). Late Pleistocene Human Skeleton and mtDNA Link Paleoamericans and Modern Native Americans. Science 344: 750-754.

2. Prufer, Kay and Matthias Meyer (2015). Comment on "Late Pleistocene Human Skeleton and mtDNA Link Paleoamericans and Modern Native Americans." Science 347:835-a.

3. Brian M. Kemp, John Lindo, Deborah A. Bolnick, Ripan S. Malhi, and James C. Chatters (2015). Response to Comment on "Late Pleistocene Human Skeleton and mtDNA Link Paleoamericans and Modern Native Americans." Science 347:835-b.

45,000-Year-Old Modern Human Genome from Western Siberia is the Earliest Yet

Svante Pääbo's team has produced the earliest Homo sapiens genome so far, extracted from a directly radiocarbon-dated femur found in western Siberia. Two AMS dates on the collagen combined to produce a calibrated date of 46,880-43,210 cal BP. Stable C and N isotope analyses indicate a diet based on C3 plants and animals that ate them. The bone is from a male whose mtDNA falls near the root of the R haplogroup, which is widespread in modern Europe and Asia. The Y chromosome is similarly ancestral a widespread Eurasian haplogroup. They called the specimen  "Ust'-Ishim" after the location in which it was recovered.

The authors find that the individual was probably related to or derived from an early population involved in the dispersal of modern humans out of Africa. Principal components analyses (PCAs) of the autosomal DNA suggest this guy was more closely related to non-Africans than Africans. In fact, when comparing the autosomal DNA only to modern non-Africans the authors observe that the specimen falls near the origin of the graph (of the first two components) and interpret that to indicate that he was equally-closely related to all modern non-Africans.

There may, however, be a better interpretation. In both PCA plots the Ust'-Ishim genome fell closest to modern Central and South Asians, which would make sense too, given the location of the find. I haven't read the Supplementary materials yet (they're 115 pages long), but I did review Section 10, which addresses this issue. To my mind, the results of the admixture analysis shown in Supplementary Figure S10.3 do indeed suggest that the Ust'-Ishim genome is closest to Central Asians, such as the Pathan, Sindhi, Burusho, Hazara, and Uygur. (The ethnic labels come from the article.) These are all central Asian peoples from Pakistan, Afghanistan, and western China. The match between the bars representing those peoples and the Ust'-Ishim genome is not perfect (the former are missing a purple patch and a bit of dark blue that appear in the ancient genome), but to me that match it does seem markedly more similar than the others. Unfortunately, the subsequent phylogenetic analyses do not address this question as they do not include central Asian genomes. So, I'm not sure I agree with the authors' interpretation that this new specimen is equally related to all Eurasians. If the Ust'-Ishim individual is more closely related to central Asians than to others, it implies a remarkable degree of geographic stability. This is an obvious hypothesis that probably should have been addressed. I hope it is investigated in the future.

They also found a 2.3 +/- 0.3% Neanderthal admixture, showing that their inter-species cuddling had already taken place by 45,000 BP. Curiously, given the date and location, there was no Denisovan admixture. The percentage of Neanderthal admixture is a bit higher than in modern humans, but not by much. As one would expect, the characteristic length of the Neanderthal DNA segments is substantially greater (~1.8-4.2 x) in the Ust'-Ishim genome than in modern ones, because the fragments were subsequently broken up further by recombination. Actually, this fragmentation process can provide an estimate of the length of time since the admixture took place, which the authors estimate as 232-430 generations earlier, or 50,000-60,000 BP, using a 29-year generation length. This corresponds approximately to some estimates of the time of expansion of modern humans out of Africa, but post-dates the putatively early modern humans from Skhul and Qafzeh. Some longer Neanderthal segments in the Ust'-Ishim genome imply that later admixture events also occurred. Interestingly, Figure 5 in the article illustrates homozygous versus heterozygous Neanderthal-derived alleles in the Ust'-Ishim genome compared to several modern ones. The Neanderthal alleles in the ancient genome are all heterozygous whereas some of the modern ones are homozygous. This makes sense, as one would expect by chance alone (panmixia) that heterozygous derived alleles would be more common earlier in the gene flow process. Probabilistically, it would take time for homozygosity to develop. Figure 5 only shows Chromosome 12, but it is suggestive.

Because the specimen is directly dated, it could be used (along with other specimens) to help calibrate the "genetic clock" that calculates age of genetic divergence from mutation rates, and the authors did so. The new estimates of mutation rates in autosomal DNA from this study are lower (slower) than some, possibly implying a longer time interval since the split from archaic human relatives.They also provided new Y Chromosome and mtDNA mutation rate estimates.

Finally, it is noteworthy that the authors find that the Ust'Ishim individual was not more closely related to the Andaman Islanders than he was to modern East Asians or Native Americans. I say this is noteworthy because some documentaries have made, in my opinion, far too much of the Andaman Islanders' genes as evidence for a coastal, southern migration route for early modern humans.

Quite the interesting article. Here's the reference:

Fu, Qiaomei, Heng Li, Priya Moorjani, Flora Jay, Sergey M. Slepchenko, Aleksei A. Bondarev, Philip L. F. Johnson, Ayinuer Aximu-Petri, Kay Prüfer, Cesare de Filippo, Matthias Meyer, Nicolas Zwyns, Domingo C. Salazar-García, Yaroslav V. Kuzmin, Susan G. Keates, Pavel A. Kosintsev, Dmitry I. Razhev, Michael P. Richards, Nikolai V. Peristov, Michael Lachmann, Katerina Douka, Thomas F. G. Higham, Montgomery Slatkin, Jean-Jacques Hublin, David Reich, Janet Kelso, T. Bence Viola & Svante Pääbo (2014). Genome sequence of a 45,000-year-old modern human from western Siberia. Nature 514: 445-450. doi:10.1038/nature13810.

Mexican DNA patterns reveal indigenous history and migrations

A major article in Science last week reveals the potential for modern DNA studies to reveal historic patterns of interaction and migration. It has implications for archaeology and especially for the use of ancient human DNA.

Moreno-Estrada, Andrés, et al. (2014). The Genetics of Mexico Recapitulates Native American Substructure and Affects Biomedical Traits. Science 344(6189): 1280-1285.

The authors sequenced large numbers (hundred of thousands) of SNPs (single nucleotide polymorphisms) from about 1000 individuals in Mexico, split between natives and mestizos (although I couldn't find in the article or the supplementary material how they determined who was which). In the data, you can see strong evidence for the genetic substructure of the population. That is, cultural, ethnic, and linguistic groups do have distinctive genetic signatures. For example, the Maya form a distinct group with additional internal structure. Within the Maya "clade" the Lacandon and Tojolobal form especially homogeneous subgroups that display little outside admixture. There's also a widespread "Central or Southern Mexican" clade (medium blue in Figure 2B and C) that impinges on the Maya. This clade is purest in southern Mexico (Oaxaca and environs), where it shows relatively little admixture. Maya genes also flowed into central Mexico (Figure 2B, bottom). In many cases, geography is a better predictor of genetics than language. Note, for example, how Nahua-speaking groups fall into more than one genetic group that is better defined spatially, which makes some sense when you think about actual mechanisms of gene flow. On the other hand, the data does contain genetic evidence of at least a couple of migrations. The most prominent one appears to go from the Yucatan Peninsula to what the authors call the Totonac region. If you plot the latitude and longitude of the sampling location given in the supplementary materials, it seems that it also close to the Huasteca. Whether this genetic contribution comes from the quite ancient Huastec migration or from a more recent Maya incursion, such as the Olmeca-Xicalanca conquest of Cacaxtla, is not addressed. I suspect, however, that the data could in theory be used as a molecular clock to date the migration, at least very roughly. Since the Huastec and Olmeca-Xicalanca migration are far apart in time, even a crude estimate would be sufficient to distinguish between them.

So, for archaeologists, there are a couple of significant facts. These kinds of studies can be used to understand prehistoric population dynamics. This is not news, perhaps. Genetic studies in Europe have been used for years to investigate the spread of agriculture and determine whether it was migration, diffusion, or both, and what the spatial dynamics were. It is interesting, however, to see it working in Mexico, which seems to me to be even more complex. Another important issue this study raises is the potential for ancient DNA to be matched to modern patterns. Presumably now we could take DNA from an archaeological individual and potentially match it preferentially to a "clade" defined by modern genetic distributions. That's an exciting possibility. It would provide far more information than the strontium isotope methods currently in use. Little ancient DNA has been sequenced in Mesoamerica, but now we have even more reason to do it.

Denisovan DNA East of Wallace's Line

As naturally and inevitably as gators eat the dogs of careless owners in Florida, the one time I carelessly leave town for personal reasons--to attend the wedding of Honduran friends in Louisiana--Science ejaculates a gob of interesting articles. There are too many of them to review in one post, so I will tackle them one at a time and hope to finish one day.

First up is this discussion in the "Perspectives" section:

Cooper, A. and C. B. Stringer (2013). Did the Denisovans Cross Wallace's Line? Science 342: 321-323.

Wallace's Line is the imaginary boundary that winds north to south through the islands of modern Indonesia marking the frontier between the mainland Asian fauna to the west and the Austronesian ecosystems to the east. Cooper and Stringer attempt to explain why small "remnant" Denisovan DNA is only found in modern human populations east of the Wallace line when the only known Denisovan lived far away in southwest Siberia. The characteristics of the Denisovan DNA sequence imply a large and diverse population in Asia. The authors, however, reject the obvious idea that most east and southeast Asian peoples once had Denisovan DNA that was later "overwritten" by subsequent gene flows. Their evidence for rejecting this interpretation is pretty strong: the DNA of isolated, remote, and evidently "indigenous" Asian populations thought to predate later migrations, such as that of the Andaman Islanders, lacks Denisovan DNA, precisely where one might most confidently expect it to be found. Moreover, recent analysis of the ancient DNA from a Chinese specimen dating from ca. 40,000 ya shows no trace of Denisovan introgression. They therefore suggest that the Denisovans who crossed the Wallace Line encountered the ancestors of Homo floresiensis, the hobbits, who were there at least 800,000 ya if not 1 mya. Perhaps, they conjecture, the Denisovan DNA signal was better preserved east of the Wallace Line because of the small size of the populations involved, which might have encouraged interbreeding, or because of selective advantages conferred by some of the genes upon rapidly advancing human populations. I didn't find their final answer to be very satisfying, but their control of the data was impressive and their discussion thought-provoking.