Genetic variation in space & time – Iceland

By Razib Khan | January 16, 2009 12:55 am

A neat new paper on Icelandic genetics, then and now, Sequences From First Settlers Reveal Rapid Evolution in Icelandic mtDNA Pool:

A major task in human genetics is to understand the nature of the evolutionary processes that have shaped the gene pools of contemporary populations. Ancient DNA studies have great potential to shed light on the evolution of populations because they provide the opportunity to sample from the same population at different points in time. Here, we show that a sample of mitochondrial DNA (mtDNA) control region sequences from 68 early medieval Icelandic skeletal remains is more closely related to sequences from contemporary inhabitants of Scotland, Ireland, and Scandinavia than to those from the modern Icelandic population. Due to a faster rate of genetic drift in the Icelandic mtDNA pool during the last 1,100 years, the sequences carried by the first settlers were better preserved in their ancestral gene pools than among their descendants in Iceland. These results demonstrate the inferential power gained in ancient DNA studies through the application of population genetics analyses to relatively large samples.

This shouldn’t be that surprising. Iceland is an island, and an island with little immigration after its first few centuries. This genetic isolation allows for the building up of differences because there is no gene flow evening out the differences in random variation which is generated by drift. This is why Iceland and Sardinia often show up as outliers on principle component maps which illustrate genetic variation between different groups.

But another issue with Iceland, as indicated in the paper, is its particular demographic history. Iceland not only has a small population, but it has historically had a small population, and one that nearly went extinct in the medieval period. This matters for what is termed in population genetics effective population size. Each generation only a particular portion of the population reproduces, so the effective population size is always smaller than the census population size (on the order of 1/3, though it varies from species to species). The effective population are those who contribute genes to the next generation. Not only do only a subset of individuals in a generation reproduce, but there is variation in this reproduction. Generally the variance can be modeled as a poisson process; some individuals contribute much, many less (poisson because of the expected value equals the variance). This disproportionate contribution has the effect of reducing effective population even further, as a subset of those who do reproduce contribute disproportionately to the next generation. Finally, the long term effective population size has modeled as a harmonic mean, which implies that the longer term effective population tends to be much closer to the lowest value to reproduce over the set of generations. This is obviously what a population bottleneck is. Once a population goes through a bottleneck a great deal of its genetic information is gone due to imperfect sampling of the pre-collapse variation in the post-collapse survivors; it takes time, either through migration or mutation, for genetic variation to be replenished.
In the case of Iceland this means that genetic drift is a very powerful force in shaping variation, specifically, in increasing the rate of extinction of haplotypes and bringing to fixation other haplotypes (haplotypes being genetic variants). Why? As population size drops, sampling variance increases. More concretely, the probability of fixation of a new mutant is:
N = population size
As you can see, the probability increases as the population size decreases. Now, remember that the deviation from expectation from generation to generation of allele frequency is:
σ = √(pq)/(2N)
σ = standard deviation
p = allele 1
q = allele 2 (and is 1 – p)
N = population size
This is what the authors mean by “faster rate of genetic drift ,” the strides that an allele can take in frequency due to the action of drift become much larger in a very small population. Isolated Iceland, buffeted by swings in population size due to inclement clime & plague during the medieval period, is a perfect test case for the power of the random walk through genetic space taking one into strange new lands.
Populations differ spatially, but they also differ in time. If you read this weblog, you know that 10,000 years ago no one had blue eyes, or could digest milk, or resist malaria. Populations change over time. Over wide swaths of northwest Eurasia on the LCT locus a new mutant swept to fixation (which confers the ability to digest milk as an adult) over the past 10,000 years. All of these peoples are now more closely related to each other than they are to any of their ancestors at that particular time depth on that gene. In fact, they are all more closely related to one particular individual in Central Eurasia who lived on the order of 10,000 years ago than they are to any of their other ancestors on that locus! This is an instance where I’m illustrating the point with selection, but the same inferences can be made about genetic changes due to neutral processes. Of course, some of the terms like “related” which intrude in from plain English start to get distorted when we move into the domain of deep time.
Because of the extreme population genetic parameters (variations in effective population size as well as isolation) which Icelanders are subject to one might argue that they are evolving faster than other European populations. But, if you read The 10,000 Year Explosion: How Civilization Accelerated Human Evolution, there are possible arguments for why Icelanders might also be evolving slower in some ways over the past 1,000 years than continental Europeans….
Note: These results were generated using mtDNA, the mitochondrial lineage, passed only from mother to daughter. This means some population genetic algebra is altered, as this is now haploid. So 1/(2N) becomes 1/N. And, because only half the population contributes to the next generation in comparison to other loci, mtDNA is more strongly effected by the vicissitudes of stochastic process compared to the autosomal genome.
Addendum: The paper also confirms previous work which suggests that over half of the mtDNA from Iceland resembles lineages in the British Isles, as opposed to Scandinavia. This can be explained by the fact that many of the Icelanders were secondary immigrants who relocated from Norse settlements in Ireland and Scotland. The mtDNA implies that many of the founding females had British or Irish foremothers. In contrast, the Y data shows that Icelanders are like Norwegians. This is a window into particular social-historical dynamics during phase of Norse domination in the British Isles.


Comments (9)

  1. Do you think that there was less than one migrant to Iceland per generation? Or perhaps the nature of long-term Ne (harmonic mean) means that the one-migrant rule is not necessarily true, e.g. maybe no migrants at the bottleneck point have an outsize effect?

  2. david,
    yeah, i think there were many times that more than 1 migrant was showing up. but there might have been generations when this didn’t happen, and those generations are also the ones that i suspect are the low points in effective population size. the think the iceland bottlenecks are probably the bigger story here than the isolation on the island. OTOH, sardinia is pretty much an outlier as well, and i don’t think it went through these bottlenecks….

  3. Sigmund

    “many of the Icelanders were secondary immigrants who relocated from Norse settlements in Ireland and Scotland”
    Or possibly female slaves taken in raids on Scotland and Ireland and traded with or taken to the Icelandic colony. I’m not sure which explanation is the most likely.
    Where is MartinR from Aardvarchaeology when we need him?

  4. I think I am following this — the better part of the Icelandic population is are from Norse men who married British women in Norse colonies in Britain?

  5. pconroy

    tdaxp said:
    the better part of the Icelandic population is are from Norse men who married British women in Norse colonies in Britain?
    These people were called Hiberno-Norse or Gallgaels.

  6. the better part of the Icelandic population is are from Norse men who married British women in Norse colonies in Britain?
    yes. the Y chromosomes show an inverted pattern; majority norse, minority british. as sigmund says the consensuality of these relationships may be debates, some of the norse did bring large retinues of slaves. but, historically in ireland there was also a process of assimilation of some irish to the norse folkways, as evidenced by the substantial minority of british & irish male lineages in iceland.

  7. bioIgnoramus

    “These people were called Hiberno-Norse or Gallgaels.” In which case it might be productive to compare the Icelanders with the population of Galloway in Scotland – which is named after the Gallgaels. That population is presumably descended from a mixture of Gallgael colonisers and indigenous Britons.

  8. Longma

    Something else to add to the mix. Iceland has undergone some serious pruning in the last few centuries, which could have added to the bottleneck effect.
    Iceland has experience two times in it’s history a 50% decline in their population…once due to disease (plague in the 14th century), the other, I believe, due to climate (18th century).

  9. John Emerson

    In Cavalli-Sforza the most genetically distinctive European populations (excluding recent immigrants) are the Icelanders, the Saami / Lapps, the Basques, and the Sardinians. In the case of Iceland it has to be the island effect intensified by several die-off bottlenecks. It’s a natural genetic experiment, given that Iceland was settled only a thousand years ago by a literate people of known origin.
    The Saami, whether they’ve always been there or whether they migrated from the East, aren’t really that surprising to anybody. The Basques confirm the idea that the Basques are survivors of a pre-Indo-European people.
    The Sardinians are partly the island effect, but it would also seem possible that the island’s isolation preserved some ancient non-Roman genes, as in the Basque country.
    The Sardinian language is no longer considered to be a dialect of Italian, though it is a romance language. Sardinia has only been under Italian jurisdiction for a century or two, and before that was either independent or ruled by one seagoing people or another. Romanization began around 238 BC. I don’t know how complete Romanization or Latinization was after 238 BC, or how long the pre-Latin language survived.On the internet the question is distorted by nationalism, but my guess is that Sardinian of today is mostly an archaic, isolated Romance language with a few words and names from the pre-Roman language.
    As an island Sardinia was both isolated, and also accessible to sea-going peoples. So there’s been relatively little genetic input by nearby nations, but there might also have been extensive input py seafarers from far away. Land migration and diffusion is different than sea diffusion and migration, in other words.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar