June 05, 2009

Geography and human adaptation (Coop et al. 2009)

From the paper:
More generally, alleles that strongly differentiate the French from both the Han and Yoruba (Figure 3D) are typically present at high frequency across all of Europe, the Middle East and South Asia (an area defined here as “west Eurasia”), and at low frequency elsewhere. This pattern of sharing across the west Eurasian populations is highly consistent with observations from random markers showing that the populations in west Eurasia form a single cluster in some analyses of worldwide population structure [40]. Allele frequencies at high- FST SNPs in two central Asian populations, the Uygur and Hazara, tend to be intermediate between west Eurasia and east Asia, consistent with observations that these populations have recent mixed ancestry between west Eurasia and east Asia [38],[40],[66].

From Figure 3: "frequency distributions for 50 of the most extreme SNPs genome-wide in the following pairs of comparisons: (A, B): SNPs for which Yoruba are highly differentiated from both Han and French; (C, D): French are differentiated from Yoruba and Han; (E, F): Han are differentiated from Yoruba and French."

From the paper:
Our results therefore suggest that local adaptation is tightly constrained by the ancestral relationships and migration rates among populations. It seems likely that selection in humans is generally not divergent enough to generate large frequency differences at individual loci between population pairs that are either recently separated, or regularly exchange migrants [53],[54]. Furthermore, populations may be too mobile, or their identities too fluid, to experience very localized pressures consistently over the several thousand years that may be required for large allele frequency changes.

However in contrast, it seems that selected alleles may not spread effectively between broad geographic regions (see Figure 3, Supplementary Figure 15 in Text S1 and [21]). Perhaps this is because populations usually adapt to similar selection pressures by parallel mutation [18],[23],[25] rather than by the spread of migrants between regions [72],[73].

UPDATE (June 5)

From the public release:
In recent years, geneticists have identified a handful of genes that have helped human populations adapt to new environments within just a few thousand years—a strikingly short timescale in evolutionary terms. However, the team found that for most genes, it can take at least 50,000-100,000 years for natural selection to spread favorable traits through a human population. According to their analysis, gene variants tend to be distributed throughout the world in patterns that reflect ancient population movements and other aspects of population history. "We don't think that selection has been strong enough to completely fine-tune the adaptation of individual human populations to their local environments," says co-author Jonathan Pritchard. "In addition to selection, demographic history -- how populations have moved around -- has exerted a strong effect on the distribution of variants."

To determine whether the frequency of a particular variant resulted from natural selection, Pritchard and his colleagues compared the distribution of variants in parts of the genome that affect the structure and regulation of proteins to the distribution of variants in parts of the genome that do not affect proteins. Since these neutral parts of the genome are less likely to be affected by natural selection, they reasoned that studying variants in these regions should reflect the demographic history of populations.

The researchers found that many previously identified genetic signals of selection may have been created by historical and demographic factors rather than by selection. When the team compared closely related populations they found few large genetic differences. If the individual populations' environments were exerting strong selective pressure, such differences should have been apparent.

Selection may still be occurring in many regions of the genome, says Pritchard. But if so, it is exerting a moderate effect on many genes that together influence a biological characteristic. "We don't know enough yet about the genetics of most human traits to be able to pick out all of the relevant variation," says Pritchard. "As functional studies go forward, people will start figuring out the phenotypes that are associated with selective signals," says lead author Graham Coop. "That will be very important, because then we can figure out what selection pressures underlie these episodes of natural selection."
Gene Expression has a long post on this study.

PLoS Genetics doi:10.1371/journal.pgen.1000500

The Role of Geography in Human Adaptation

Graham Coop et al.


Various observations argue for a role of adaptation in recent human evolution, including results from genome-wide studies and analyses of selection signals at candidate genes. Here, we use genome-wide SNP data from the HapMap and CEPH-Human Genome Diversity Panel samples to study the geographic distributions of putatively selected alleles at a range of geographic scales. We find that the average allele frequency divergence is highly predictive of the most extreme FST values across the whole genome. On a broad scale, the geographic distribution of putatively selected alleles almost invariably conforms to population clusters identified using randomly chosen genetic markers. Given this structure, there are surprisingly few fixed or nearly fixed differences between human populations. Among the nearly fixed differences that do exist, nearly all are due to fixation events that occurred outside of Africa, and most appear in East Asia. These patterns suggest that selection is often weak enough that neutral processes—especially population history, migration, and drift—exert powerful influences over the fate and geographic distribution of selected alleles.



Maju said...

I have written more extensely at my blog on this paper. Just two brief comments:

1. West and East Eurasians appear clearly as distinct (within the Eurasian supergroup and within the limited variability of human genetics). They also appear as distinct in pigmentation startegies. Overall this suggests that the coastal migration model is correct and that different founder effects happened in South Asia and SE Asia, including pigmentation genes.

2. Selection is rather weak anyhow and chaotic accidents (founder effects, drift) instead are more important, even in the pre-selection of what moderate fitness selection would eventually make dominant.

terryt said...

"gene variants tend to be distributed throughout the world in patterns that reflect ancient population movements and other aspects of population history".

And that seems to be so for the geographic variation of most species, so I'm not surprised it has been important for human genetic variation.