What conclusions can we draw from Neanderthal DNA pt.2

ResearchBlogging.org4. Nuclear DNA: Forays into 3 billion base pairs

4.1 Before Vi-80

The Vindija-80 (Vi-80) specimen is an important find for geneticists: it yielded a minimally contaminated sample and provided those first steps into Neanderthal genomics.

Previously, attempts at retrieving ancient nuclear DNA sequences proved to be a notoriously difficult process, plagued with problems of degradation, contamination and chemical damage (Hofreiter et al., 2001). Researchers also need to contend with quantities of nuclear genome available: for every nuclear genome there are approximately several hundred mtDNAs (Green et al., 2008). The severity of these problems, especially contamination, is magnified through Neanderthal genetic similarity with humans (Green et al., 2006). This is troubling because nuclear DNA presents far less variability than mtDNA (Russell, 2002). As a result, huge stretches of nuclear sequences are required to find a significant number of polymorphisms (ibid). Such implications meant that discovering endogenous DNA sequences requires sifting through a large corpus of “[…] more than 70 Neanderthal bone and tooth samples from different sites in Europe and western Asia” (Green et al., 2006, pg. 331).

To make Vi-80 even more poignant, its sequencing was performed from two different perspectives: Green et al. (ibid) use the previously mentioned 454 method, while the Rubin Group (Noonan et al., 2006) developed a metagenomic library and subjected it to high-throughput sequencing.

Arguably, the important distinction between these two methods is merely that 454 sequencing is considerably faster than labouring over library construction (ibid). This explains the huge differences in sample size: the 454 group sequencing produced approximately one million base pairs (Green et al., 2006), compared with the 65,250 base pairs (Noonan et al., 2006) used by the Rubin group. However, Noonan et al.’s approach apparently offers a superior method of searching for specific genetic loci, important in the search for single nucleotide polymorphisms. But for the moment, the advantages of this slower approach lie in its ability for a greater degree of accuracy against contamination and other errors (ibid).

4.2 The search for derived nucleotides

The data from Vi-80 is for the most part inconclusive. For instance, both teams arrive at dates estimating genetic divergence at ~706,000 years ago (Noonan et al., 2006) and ~516,000 years ago (Green et al., 2006) respectively. On the surface these dates are in concordance with fossil record data and previous mtDNA studies; keeping with the propensity of a genetic distinctiveness between modern humans and Neanderthals (Relethford, 2008). Yet the obvious discrepancy between dates does not make sense, considering both teams used the same sample (Wall and Kim, 2007).

Now, using their more recent date the 454 group utilise a broad comparative approach between modern human[1], chimpanzee and Neanderthal genomes (Green et al., 2006). Searching for SNPs in particular can shed light on genetic similarities and differences by distinguishing between ancestral nucleotides (those found on Chimpanzees) and derived nucleotides (more recent and unique to humans) (Relethford, 2008). By cataloguing these derived single base pair changes in humans, and then comparing them to the Neanderthal genome data, Green et al. (2006) found that the Neanderthal sequence “[…] carries the derived allele for roughly 30% of human SNPs in the public domain.” (Wall and Kim, 2007, pg. 1862).

Finding a significant amount of derived alleles[2] in the Neanderthal genome is discordant with population split models. It is also suggestive of Neanderthal-human admix, with Green et al. adding that “[…] gene flow may have occurred predominantly from modern human males into Neanderthals.” (ibid, pg. 335)[3].

Meanwhile, Noonan et al. (2006) opt for a different approach and focus on derived alleles in the modern European gene pool matching derived alleles found in Vi-80. Their dataset produced a 0% (CI: 0%-20%[4]) likelihood of contribution to the European gene pool from Neanderthal ancestry (ibid). Furthermore, the Neanderthal sequence matched Chimpanzee DNA at a variety of locations where Europeans did not; hence the older divergence time, as Neanderthals share more ancestral nucleotides with Chimpanzees (ibid).

Resolving these discrepancies is still open to interpretation, mostly depending upon what side of the admixture hypotheses you agree with. Some suggest variations in the proportions of Vi-80 genome sequenced are enough to account for these differences (Relethford, 2008). A peer review by Wall and Kim (2007) suggest a methodological fault is to blame, with emphasis placed on Green et al.’s study. An additional review of the data also confirms this notion of high contamination levels (Green et al., 2008).

To complicate matters further, Forhan et al. (2008) reanalysed some of the data from Noonan et al.’s (2007) study and found the initial rejection of admixture was slightly forthcoming, “[…] if the population size of Homo sapiens was comparable to the Neanderthal population size.” (Forhan et al., 2008, pg. 71).

Until the inconsistencies between both groups are resolved[5] the reliability of their studies remains difficult to ascertain. For now, our most current understanding of Neanderthal nuclear DNA comes from a different approach.

Unlike the genome sequencing performed on Vi-80, whereby vast swathes of the genome are compared, it is possible to narrow the search on specific genetic loci where a positive selective sweep appears to have taken place (Hawks et al., 2007). Several candidate genes have been proposed, specifically in areas associated with brain development (Sabeti et al., 2006) and pigmentation morphology (Williamson et al., 2007). Seeking out these specific haplogroups, and examining their SNPs, is the basis of two recent studies into the Neanderthal genome: FOXP2 (Krause et al., 2007), a gene involved in brain development and MC1R (Lalueza-Fox et al., 2007), a gene involved in pigmentation.

4.2 FOXP2: Evidence of introgression?

Since discovering its implications in language-related disorders, such as orofacial dyspraxia and dysphasia (Li et al., 2007), forkhead box P2 (FOXP2) has become a widely discussed gene. Found on chromosome 7, FOXP2 is highly conserved across a broad variety of species, showing only two adaptive amino acid changes between the human and chimpanzee variants (Enard et al., 2002); suggesting it played a crucial role in the development of modern humans. When coupled with studies in mice and echolocating bats (cf. Li et al., 2007), the consensus in regards to the wider functional role of FoxP2[6] seems to support regulatory “[…] function[s] in sensorimotor integration and motor learning” (ibid, pg. 1).

FOXP2 then, seems to be an obvious candidate in the search for specific human haplotypes in Neanderthals. Particularly, its conserved status suggests a recent selective sweep on those two amino acid changes (Krause et al., 2007). With this in mind, it is easily understood why discovering the human FOXP2 variant in a Neanderthal specimen would have profound implications. First, though, it is essential to establish what FOXP2 cannot tell us about Neanderthals.

It seems probable Neanderthals had the aptitude for speech. Whether or not they actually had language, or even a protolanguage[7], is something far too expansive to be suitably addressed in this essay[8]. Secondly, as the regulatory pathway of FOXP2 in expression is still not entirely understood in humans (Szathmáry and Számadó, 2008)[9], it is inaccurate to proclaim the two polymorphisms in FOXP2 account for the inherent complexity found within language (ibid).

As for what it can tell us: using two separate Neanderthal specimens, Krause et al. (2007) claim a fixation point for FOXP2 of 300,000 BP or older. Not only does this date place them at odds with other findings, which suggest FOXP2 fixation occurred within the last 200,000 years (Enard et al., 2002), it also predates the predicted divergence times between modern human and Neanderthal populations from their common ancestor (Krause et al., 2007). Assuming this interpretation is correct, the appearance of FOXP2 need not invoke the possibility of admixture.

Additionally, Krause et al. use the Y-Chromosome data to support these assertions. It seems two of the specimens do not have a human-specific derived SNP, which according to the authors “[…] shows that neither the maternally inherited mtDNA nor the paternally inherited Y chromosome shows evidence of gene flow from modern humans into Neanderthals” (ibid, pg. 2).

A puzzling aspect surrounding the 300,000 BP fixation date is that current methods of genomic sequencing can only detect positive selection up to ~250,000 BP (Sabeti et al., 2006). Equally as perplexed by this fixation date are Coop et al. (2008), who offer two alternative explanations[10] for the data: either the Neanderthal specimens were contaminated, or admixture took place with modern humans.

Ignoring contamination qualms, Coop et al. (2008) use an adaptive introgression model to propose a more recent selective sweep of ~42,000 BP. In this scenario an archaic allele can, through relatively rare interbreeding episodes[11], spread throughout a rapidly expanding population by retaining an adaptive advantage (Hawks and Cochran, 2006). This is especially relevant to Neanderthal-human admixture: “Because hybridizing species share a large proportion of their genetic background, a new allele that is adaptive in one species may retain its selective advantage after introgressing into another.” (ibid, pg. 103).

As one might expect, searching for evidence of adaptive introgression in extant species is difficult enough (Castric et al., 2008), without adding extinct populations into the mix. Though considering the conserved variation of FOXP2, and its surrounding regions in most mammalian species[12], positive selection probably did take place recently (Kelly and Swanson, 2008). However, if selection happened earlier than previously thought, as suggested by Krause et al. (2007), then the linked derived alleles of FOXP2 should show evidence of more dispersal via genetic drift – and this appears not to be the case in modern human evolution (Coop et al., 2008).

Accounting for the Y-Chromosome data is slightly trickier, but not necessarily impossible. If the modern human Y-chromosome was subjected to selection, and only recently became fixed[13] in a rapidly expanding population, then a different story emerges: the Neanderthal Y-chromosome is expected not show polymorphisms that emerged later in human ancestry (Hawks et al., 2007). In other words, the Y-Chromosome data can be interpreted as either for or against admixture.

4.3 MC1R: Did the Neanderthals have red hair?

For all the speculation surrounding FOXP2 there is little certainty of its exact regulatory functions. Melanocortin 1 receptor (mc1r) meanwhile is the exact opposite: not only do we know that the human variant of MC1R is involved in pigmentation, we can also perform a functional analysis to determine its phenotypic properties (Lalueza-Fox et al., 2007).

The MC1R protein in humans helps maintain a certain continuity of pigments found in hair and skin. In most cases humans express one of these two pigments: eumelanin and pheomelanin. Whilst eumelanin is found in the vast majority, phenomelanin is far rarer and is usually responsible for fairer skin, freckles and red hair. This is because the usual process of MC1R gene enables the coversion of pheomelanin into eumelanin, except when it is broken. Here, the partial loss of function in a homozygous state allows phenomelanin to build up unchecked in melanocytes[14] and is responsible for the characteristics described above (cf. Tennen, 2007).

Again, using two Neanderthal specimens from Monti Lessi (Italy) and El Sidrón (Spain) Lalueza-Fox et al. (2007) employ the PCR method to extract and amplify a 128-bp fragment of mc1r. Dubbed the R307G substitution, this fragment is not found anywhere in current human sequences and implies the sample is genuinely endogenous to Neanderthals (ibid) – something the FOXP2 experiment could not confirm with the same degree of authority.

Now having a Neanderthal-derived mc1r variant, Lalueza-Fox et al. then perform a functional analysis. They find the extracellular signalling of R307G operates in a similar manner the MC1R gene variant found in humans; producing an increase in the pheomelanin pigmentation. This highly suggests Neanderthals had some red-hair, light skinned individuals. As we cannot determine if the sequenced individuals were homozygous or heterozygous for this allele, then calculating the number of Neanderthals with these traits is unlikely. A conservative estimate places the homozygous allele being present in around 1% of the total Neanderthal population (ibid).

Extrapolating these results into any meaningful discussion about selection in Neanderthals is for the most part negated. Though the authors do speculate in their discussion that these results “once more raise the question of whether reduced pigmentation may have been advantageous in Europe, for example via ultraviolet-light mediated vitamin D synthesis” (ibid, pg. 1454). Still, unless we can discern the level at which the R307G mutation occurred in the Neanderthal population, hypotheses about the mc1r allele are best explained through genetic drift and purifying selection.

5. Conclusion

Although the Neanderthal DNA debate fervently rages on, there are some areas where contention is being eroded into a vague semblance of accordance. For instance, it seems likely much of our modern human ancestry comes from Africa. This is largely inferred using mtDNA, which shows little evidence for any significant genetic contribution from Neanderthals. Furthermore, Neanderthals might be considered a separate species, as their mtDNA is approximately three times as different as human-human mtDNA variation. Nuclear DNA also offers some support here, in that Neanderthals contain alleles, such as mc1r, not found anywhere in current human populations. Overall, it leads to a picture of minimal gene flow from Neanderthals into humans, reflected in the inherent genetic discrepancies between the two.

This is not a complete negation of a multiregional perspective of human origins. Our genetic legacy may be primarily of archaic African origin, but it still does not filter out the possibility of key genetic contributions taking place from Neanderthals. Given their genetic similarities, and that human and Neanderthal populations geographically overlapped, gene flow is a very probably scenario. FOXP2 is our best evidence of this so far, having likely developed after the human-Neanderthal ancestral split but still being present in both populations. Did it introgress because of a particular adaptive value, namely in brain development? If so, this is strongest indication of admixture between Neanderthals and humans. However, discerning which group it introgressed from (Neanderthal-to-humans or humans-to-Neanderthals) is currently beyond explanation. It means we are still lacking a fundamental component in recognising if any genes of adaptive value can be attributed to a Neanderthal lineage.

Moving outside the human origins debate, there are other conclusions we can yield from Neanderthal DNA. First, it appears the mtDNA genepool of Neanderthals is very restricted, even more so than humans. By implication, the reduced variety in genetic differences between individual Neanderthals may mean the Neanderthal population size was small. Secondly, thanks to the mc1r study, new avenues are on the horizon in regards to understanding Neanderthal phenotypic properties.

Having a picture of the complete Neanderthal nuclear genome is the next necessary step to enhance our understanding. First and foremost it will hopefully allay the differences found in the Vi-80 studies. Also, some contentions over the significance of any non-African archaic genetic endowments might be solved. In particular, we will see if there is a correlation in both humans and Neanderthals for any genes that underwent positive selection after separating from their common ancestor. Lastly, the limitations of both mtDNA and nuclear DNA studies tell us our understanding needs to be multidisciplinary in its composition. Only then can a consensus begin emerge as to the fate of the Neanderthals.

[1] Green et al. (2006) used two genome data sets for their comparison: 786 SNPs from HapMap and 318 SNPs from Perlegen.

[2] Derived allele and derived nucleotide are used interchangeably.

[3] Allowing them to make this assumption is that Vi-80 contained Y-Chromosome data, meaning it came from a Neanderthal male.

[4] The confidence interval provided here does leave open the interpretation for a maximum of 20% admixture. Though if considered contextually, it is more suggestive of Neanderthal extinction.

[5] This resolution will hopefully come from the Neanderthal genome sequence due at the end of this year. For more information: http://www.newscientist.com/article/dn16224-neanderthal-genome-already-giving-up-its-secrets.html

[6] From here on in, FOXP2 will imply the derived human variant, while FoxP2 is the variant within all other species.

[7] See The Symbolic Species by Terrance Deacon (1998) for a good account of the protolanguage hypothesis.

[8] There are still ongoing anatomical debates over the Neanderthal hyoid bones, with some suggesting a human-like anatomy (Wolpoff, 2004) while others suggest they could not produce quantal sounds (Liberman, 2007).

[9] Fisher et al. (2008) very recently discovered some regulatory function of FOXP2 by analysing CNTNAP2 polymorphisms found in children with specific language impairments.

[10] There is a third explanation: a linked locus of FOXP2 was the target of selection, rather than FOXP2 itself (Coop et al., 2008). Though this is more of a side note in their study, and has too many ramifications to flesh out in this essay.

[11] Coop et al. (2008) actually argue for “just a few viable hybrids” (pg. 1257). Though the exact levels of admixture required, for adaptive introgession to be a viable explanation, vary from model to model (Cyran and Kimmel, 2005).

[12] Echolocation in bats demonstrates a high degree of variability on the FoxP2 site according to the Li et al paper.

[13] Some suggestions place Y-Chromosome fixation as early as ~50,000 BP (Brookfield, 2000).

[14] Melanocytes are skin cells normally found at the lower part of the epidermis (Tennen, 2007).

Main References
Green, R., Krause, J., Ptak, S., Briggs, A., Ronan, M., Simons, J., Du, L., Egholm, M., Rothberg, J., Paunovic, M., & Pääbo, S. (2006). Analysis of one million base pairs of Neanderthal DNA Nature, 444 (7117), 330-336 DOI: 10.1038/nature05336

Briggs AW, Good JM, Green RE, Krause J, Maricic T, Stenzel U, Lalueza-Fox C, Rudan P, Brajkovic D, Kucan Z, Gusic I, Schmitz R, Doronichev VB, Golovanova LV, de la Rasilla M, Fortea J, Rosas A, & Pääbo S (2009). Targeted retrieval and analysis of five Neandertal mtDNA genomes. Science (New York, N.Y.), 325 (5938), 318-21 PMID: 19608918

Krause J, Lalueza-Fox C, Orlando L, Enard W, Green RE, Burbano HA, Hublin JJ, Hänni C, Fortea J, de la Rasilla M, Bertranpetit J, Rosas A, & Pääbo S (2007). The derived FOXP2 variant of modern humans was shared with Neandertals. Current biology : CB, 17 (21), 1908-12 PMID: 17949978

Lalueza-Fox C, Römpler H, Caramelli D, Stäubert C, Catalano G, Hughes D, Rohland N, Pilli E, Longo L, Condemi S, de la Rasilla M, Fortea J, Rosas A, Stoneking M, Schöneberg T, Bertranpetit J, & Hofreiter M (2007). A melanocortin 1 receptor allele suggests varying pigmentation among Neanderthals. Science (New York, N.Y.), 318 (5855), 1453-5 PMID: 17962522

Lalueza-Fox C, Römpler H, Caramelli D, Stäubert C, Catalano G, Hughes D, Rohland N, Pilli E, Longo L, Condemi S, de la Rasilla M, Fortea J, Rosas A, Stoneking M, Schöneberg T, Bertranpetit J, & Hofreiter M (2007). A melanocortin 1 receptor allele suggests varying pigmentation among Neanderthals. Science (New York, N.Y.), 318 (5855), 1453-5 PMID: 17962522

Coop, G., Bullaughey, K., Luca, F., & Przeworski, M. (2008). The Timing of Selection at the Human FOXP2 Gene Molecular Biology and Evolution, 25 (7), 1257-1259 DOI: 10.1093/molbev/msn091

John Hawks’ weblog is an indispensable resource for Neanderthal genetics: http://johnhawks.net/weblog/reviews/neandertal_dna