1. ¹Human Genetics Unit, Medical Research Council, Western General Hospital, Edinburgh, Scotland, UK
2. ²Andrija Stampar School of Public Health, Faculty of Medicine, University of Zagreb, Zagreb, Croatia
3. ³Institute for Anthropological Research, Zagreb, Croatia
4. ⁴Department of Public Health Sciences, University of Edinburgh Medical School, Edinburgh, Scotland, UK

Correspondence: Dr I Rudan, Department of Medical Statistics, Epidemiology and Medical Informatics, Andrija Stampar School of Public Health, Faculty of Medicine, University of Zagreb, Rockefellerova 4, 10000 Zagreb, Croatia. Tel: +385 91 3011 575; Fax: +385 91 2903 752; E-mail: irudan@mef.hr

⁵These authors contributed equally to this work.

Received 17 October 2005; Revised 21 December 2005; Accepted 10 January 2006; Published online 22 February 2006.

Abstract

Communities with increased shared ancestry represent invaluable tools for genetic studies of complex traits. '1001 Dalmatians' research program collects biomedical information for genetic epidemiological research from multiple small isolated populations ('metapopulation') in the islands of Dalmatia, Croatia. Random samples of 100 individuals from 10 small island settlements (n<2000 inhabitants) were collected in 2002 and 2003. These island communities were carefully chosen to represent a wide range of distinct and well-documented demographic histories. Here, we analysed their genetic make-up using 26 short tandem repeat (STR) markers, at least 5 cM apart. We found a very high level of differentiation between most of these island communities based on Wright's fixation indexes, even within the same island. The model-based clustering algorithm, implemented in STRUCTURE, defined six clusters with very distinct genetic signatures, four of which corresponded to single villages. The extent of background LD, assessed with eight linked markers on Xq13-21, paralleled the extent of differentiation and was also very high in most of the populations under study. For each population, demographic history was characterised and 12 'demographic history' variables were tentatively defined. Following stepwise regression, the demographic history variable that most significantly predicted the extent of LD was the proportion of locally born grandparents. Strong isolation and endogamy are likely to be the main forces maintaining this highly structured overall population.