July 28, 2009

Ancestry informative marker set for determining continental origin (Nassir et al. 2009)

UPDATE: This PCA plot from the paper is very instructive:




In PC1 vs PC2 (top left), the five major races (Caucasoids, East Asian Mongoloids, Amerindians, Australoids, Negroids) are clearly separable, but their relationships are not very clear. Australoids are somewhere between Caucasoids and East Asians. If we were to go by this figure alone, we might even infer that they are some type of mix of the two. But, look at PC3 vs PC4 (top right). Here it is the case that Australoids are not intermediate between Caucasoids and Mongoloids, but occupy their own space, and we can conclude that they are not in fact such a mix.

Compare with Mexicans (bottom left), who occupy the same space as East Asians in PC1 vs PC2. Does that mean that they are East Asians? An alternative explanation is that they are a mix of Caucasoids and Amerindians, since they occupy an intermediate position between the two groups, which happens to coincide with the position of East Asians. But, if we look at PC3 vs P4 (bottom right), it is clear that Mexicans are indeed better explained as a Caucasoid-Amerindian mix, as they occupy an intermediate position between Caucasoids and Columbians/Quechuans (who are on the left), and not at all that of East Asians (who are on the right).

Next, consider African Americans. In PC1 vs PC2 they occupy an intermediate position between Negroids and Caucasoids (bottom left), as expected by their known ancestry. But, what if we didn't know about their history, and we wanted to infer their origin based on their genomes, as we did with Australoids? Unlike Australoids, in PC3 vs PC4, African Americans do not form a cluster of their own, but overlap with both Caucasoids and Negroids who occupy the same space in these two dimensions. Thus, we are more certain that they are indeed a Caucasoid-Negroid mixed population.

Next, consider Mozabites and South Asians, both of which deviate from Caucasoids, in a Negroid, and Mongoloid direction respectively (top left). North African Mozabites may indeed by Negroid-influenced, as is evident in the STRUCTURE analysis of this paper. South Asian Indians, however, who show "admixture" with the East Asian cluster at K=5 in the STRUCTURE analysis, are revealed to form a cluster of their own at higher K (pink), suggesting that they are not, in fact a Caucasoid-Mongoloid mixed population.

The important lesson from all of this, is to use as much information as possible when trying to examine population relationships from graphical plots, because things may not always be "what they seem", and a number of alternative explanations may result in identical two-dimensional plots.

BMC Genetics 2009, 10:39doi:10.1186/1471-2156-10-39

An ancestry informative marker set for determining continental origin: validation and extension using human genome diversity panels

Rami Nassir et al.

Abstract (provisional)

Background

Case-control genetic studies of complex human diseases can be confounded by population stratification. This issue can be addressed using panels of ancestry informative markers (AIMs) that can provide substantial population substructure information. Previously, we described a panel of 128 SNP AIMs that were designed as a tool for ascertaining the origins of subjects from Europe, Sub-Saharan Africa, Americas, and East Asia.

Results

In this study, genotypes from Human Genome Diversity Panel populations were used to further evaluate a 93 SNP AIM panel, a subset of the 128 AIMS set, for distinguishing continental origins. Using both model-based and relatively model-independent methods, we here confirm the ability of this AIM set to distinguish diverse population groups that were not previously evaluated. This study included multiple population groups from Oceana, South Asia, East Asia, Sub-Saharan Africa, North and South America, and Europe. In addition, the 93 AIM set provides population substructure information that can, for example, distinguish Arab and Ashkenazi from Northern European population groups and Pygmy from other Sub-Saharan African population groups.

Conclusion

These data provide additional support for using the 93 AIM set to efficiently identify continental subject groups for genetic studies, to identify study population outliers, and to control for admixture in association studies.

Link

4 comments:

eurologist said...

Australoids are somewhere between Caucasoids and East Asians. If we were to go by this figure alone, we might even infer that they are some type of mix of the two. But, look at PC3 vs PC4 (top right). Here it is the case that Australoids are not intermediate between Caucasoids and Mongoloids, but occupy their own space, and we can conclude that they are not in fact such a mix.

Another way of saying this is that like Indians (South Asians), Australoids don't have any of the main distinguishing features that Africans (PC1), or Amerindians, East Asians, and Europeans have (PC2). Yet, they have the same set that distinguishes East Asians and Indians from the most of the remainder in PC3, and on top of that, have their very own set in PC4.

Also, they are a bit less non-African than most. From all that, it seems clear that their origin is somewhat ancient but shares much with both South and East Asia.

Finally, unlike PC3, I think PC2 does not indicate a mixed South and East Asian background, because there is very little overlap. The lack of PC2 distinction more likely shows the lack of changes that have accumulated in other Asians, since their divergence.

onur said...

I think Greeks, Turks, Cypriots, Armenians, Azeris, all Caucasian, Iranic, Central Asian Turkic (including Altaians and Iranian Turkmens), Afro-Asiatic and even Nilo-Saharan populations should be included in these kinds of analyses, as these people are at a bridge position between very different regions of the world and may show unique mixtures and characteristics depending on the population.

onur said...

Additionally, all Uralic populations should be included in such analyses.

onur said...

And lastly, Russian and Ukrainian Turkic populations like Tatars.