December 19, 2004
A major study by deCODE scientists of the geographic distribution of genetic variability in the Icelandic population has been published in the online edition of Nature Genetics. The paper by Agnar Helgason, et al., is entitled “An Icelandic example of the impact of population structure on association studies,” and will appear in the January print edition of the journal.
The authors used deCODE’s nationwide genealogical database to trace the geographical roots of all Icelanders born since 1850, divided into 30-year birth cohorts. The results demonstrate that Icelanders have tended to live in the same part of the country as did their forbears even five generations back. This tendency is most evident among the oldest group analyzed — those born between 1850-1875. The effect of urbanization and increased mixing within the population is clearly seen in more recent decades, but the trend is still evident in the youngest group, born between 1970 and 1995.
Dr. Helgason’s team then examined what effect this historical and geographical phenomenon might have on the genetic structure of the population; that is, whether the frequency of certain alleles, or versions, of genetic markers would tend to vary between geographical regions. To answer this question the authors measured variability in 40 genetic markers across the genome in more than 43,000 Icelanders, and analyzed this data in the context of geographical origins. Indeed, in the oldest cohort, which includes individuals born between 1895 and 1935, the allele frequencies of all 40 markers differed significantly according to region. As expected, this effect decreased progressively in the more recent cohorts, although even for those born between 1960 and 2000, allele frequencies continued to exhibit variability dependent upon place of birth.
These findings are highly relevant to medical genetics research because they establish, even in a small and relatively homogeneous population like that of Iceland, that there exist subpopulations with discernible genetic differences. The authors note that the effect of such stratification in larger, more diverse populations like those of Europe and the United States will be even greater, underscoring the need to take population structure into account particularly in the design of large-scale association studies to correlate genetic variation with susceptibility to disease. Ensuring that patient and control cohorts are in as far as possible drawn from the same groups, the authors conclude, will aid in maximizing the chance of finding such associations when they exist and minimizing the risk of false positive results.
deCODE’s unique genealogy database makes this kind of assessment possible. This database contains information on the genealogies of the entire present day population of Iceland, more than 95% of Icelanders born after 1700, and stretches back to the settlement of the country around 900 AD. deCODE also has gathered detailed genotypic and medical data from more than 100,000 volunteer participants in its gene research in Iceland – over half of the adult population.