1. ¹Department of Biotechnology, Institute of Molecular and Cell Biology, University of Tartu, Estonia
2. ²Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany
3. ³Institute of Biochemistry and Genetics, Ufa Research Center, Russian Academy of Sciences, Ufa 450054, Russia
4. ⁴Department of Bioinformatics, Institute of Molecular and Cell Biology, University of Tartu, Estonia

Correspondence: Professor M Laan, Department of Biotechnology, Institute of Molecular and Cell Biology, University of Tartu; Riia St 23, 51010 Tartu, Estonia. Tel: +372 7 375008; fax: +372 7 420286; E-mail: maris@ebc.ee

Received 4 May 2004; Revised 19 October 2004; Accepted 26 October 2004; Published online 22 December 2004.

Abstract

Linkage disequilibrium (LD) structure is still unpredictable because the interplay of regional recombination rate and demographic history is poorly understood. We have compared the distribution of LD across two genomic regions differing in crossing-over activity – Xq13 (0.166 cM/Mb) and Xp22 (1.3 cM/Mb) – in 15 Eurasian populations. Demographic events predicted to increase the LD level – genetic drift, bottleneck and admixture – had a very strong impact on extent and patterns of regional LD across Xq13 compared to Xp22. The haplotype distribution of the DXS1225–DXS8082 microsatellites from Xq13 exhibiting strong association in all populations was remarkably influenced by population history. European populations shared one common haplotype with a frequency of 25–40%. The Volga-Ural populations studied, living at the geographic borderline of Europe, showed elevated LD as well as harboring a significant fraction of haplotypes originating from East Asia, thus reflecting their past migrations and admixture. In the young Kuusamo isolate from Finland, a bottleneck has led to allelic associations between loci and shifted the haplotype distribution, but has much less affected single microsatellite allele frequencies compared to the main Finnish population. The data show that the footprint of a demographic event is longer preserved in haplotype distribution within a region of low crossing-over rate, than in the information content of a single marker, or between actively recombining markers. As the knowledge of LD patterns is often chosen to assist association mapping of common disease, our conclusions emphasize the importance of understanding the history, structure and variation of a study population.