Commentary on Sokal et al.'s "Historical Population Movements in Europe Influence Genetic Relationships in Modern Samples" (1996)

Correction:

Due to an error at the printer, the original article page range (605–606) was incorrect. The correct pagination is 607–608 and is reflected in this updated article. Click here for the corrected PDF.

The genetic history of Europe has been the focus of many studies looking at the relative influence of geography, language, and prehistoric and historic population movements on genetic differences between populations. Two groups of researchers have had a major impact on the evolution of these studies—Luca Cavalli-Sforza and colleagues (e.g., Cavalli-Sforza et al. 1994; Menozzi et al. 1978) and Robert Sokal and colleagues. The paper reprinted here is one of many of Sokal's innovative approaches to population genetics, adroitly integrating anthropological data and hypotheses and genetic data in a rigorous and statistically precise manner.

The focus of the paper is the use of ethnohistorical data to test the hypotheses of the influence of historical population movements over the past 4,000 years on the genetic structure of contemporary human populations across Europe. Sokal et al. used ethnohistorical data to construct a distance measure based on the likely relative ancestral contribution of different ethnic groups, which were defined in terms of language family membership. They found a significant correlation with genetic distance that was derived from a large sample of classical genetic markers. Further, they showed that this correlation remained even after controlling for geographic distance, thus showing that the expected relationship was not a byproduct of spatial autocorrelation. In addition, they found that the strength of the correlations increased over time, particularly during the past 400 years.

This paper is an excellent example of the multidisciplinary approaches that Bob Sokal brought to investigations of genetic history. Ethnohistory is a key factor in the studies of anthropological genetics, because the origin of different groups and their contact with other populations, both near and far, affect genetic variation. Most often, the history of different populations is used to make indirect inferences regarding observed patterns of genetic similarity. For example, in my study of 19th-century Irish anthropometrics, the observed distinction of two island populations was explained by noting that their history pointed to a large influx of English soldiers (Relethford 1988). For many studies, historical hypotheses are treated in such an indirect fashion. What makes Sokal's paper so innovative is that he and his colleagues quantified European ethnohistory through the compilation of a massive database on European-population movements (this database is still available at the Web page cited in the paper). The database alone is a remarkable achievement that represents the enormous potential in translating anthropological hypotheses into testable statistical hypotheses. Whereas the complexity of quantifying historic population movements from ethnohistorical data might appear daunting to many (myself included), Sokal and colleagues presented their approach clearly and logically, answering potential criticisms and noting both strengths and weaknesses of their approach.

It has been over 16 years since this paper was published, but its results and methodology still hold up. The most dramatic change over the intervening years has been the increasing use of DNA markers, providing both new analytic approaches and an enormous number of loci. For example, Novembre et al.'s (2008) analysis of genetic variation in Europeans used more than half a million SNPs, orders of magnitude above the 26 genetic systems used in Sokal's paper. Developments in genetic technology have also allowed the increasing use of ancient DNA analysis to help probe questions of prehistoric movements into Europe (e.g., Haak et al. 2010; Skoglund et al. 2012). It is likely that such developments will continue as the new age of genomic analysis continues to unfold. I think that these newer studies will supplement, though not replace, the earlier studies of European genetic history. The methods outlined by Sokal and colleagues and their massive achievement in quantifying their ethnohistorical database should continue to guide future analysis. At present, DNA markers still do not cover the number of populations or degree of geographic sampling that have been obtained from previous decades of...