Clarifying Bauchet et al. (2007)

February 22, 2011

This was among the first in the recent spate of genome-wide population structure studies, and better ones have been done since then that paint a clearer picture. Here, the sampled populations are divided into broad "northern" and "southern" clusters, and the authors conclude the following about Italians:

Conversely, Italy appears to be a zone of sharp differentiation over small distances. Some Italians cluster with the northern Europeans, whereas others fall into the southeastern grouping (fig. 4A).

Even though the text says no such thing, everyone automatically assumes that it's Northern Italians who are clustering with Northern Europeans, and that the observed pattern is representative of Italy's genetic structure. But in fact, the three individuals in question are Southern Italians, and the North-Central Italians (two Tuscans from the Coriell database) cluster even farther to the "south" than Sicilians:


Either way, the authors' conclusion is unfounded, because those few individuals are outliers that don't represent the average. In subsequent studies that have used larger sample populations and more genetic markers, all Italians cluster with other Southern Europeans according to geographical location, and are clearly distinguished from both Northern Europeans and Ashkenazi Jews. Even in a similar study from the year before, Seldin et al. (2006), there was no overlap between a sample of 86 Northern, Central and Southern Italians and various samples from farther north in Europe.

Although it is a bit surprising to find so many outliers in such a small sample, they're not unusual in and of themselves, and more often than not they're Northern and Central Europeans. In fact, in Bauchet's Figure 4A, you can see a German individual in the "southern" cluster, and different studies have shown similar outliers from Britain, France, Belgium, Scandinavia and Austria, among other places.

On a side note, it's also worth mentioning that non-Caucasoid admixture from Africa (Negroid) and Asia (Mongoloid) is as negligible in Bauchet's Italian samples as it is in the other European samples.

---------------
Bauchet et al. "Measuring European Population Stratification with Microarray Genotype Data". Am J Hum Genet, 2007.

0 comments