Human mitochondrial DNA variation and the origin of Basques

Abstract

The hypervariable segment I of the control region of the mtDNA (positions 16024–16383) was PCR-amplified from mouth scrape and hairs and sequenced in 45 unrelated individuals of pure matrilineal Basque descent. Twenty-seven different sequences were found, of which 21 are unique to the Basques. The allelic partition observed, together with resampling experiments, suggested that much more variation remained to be discovered. The mean pairwise difference in number of nucleotides between individuals was 3·15, a very low value. Moreover, the number of steps for the most parsimonious tree is very low compared to the number of different sequences. Both findings suggest that the Basque population was founded by a few lineages that diverged in a short time span. The number of nucleotide differences between individuals was shown not to be influenced by the distance between their birthplaces, thus validating the sampling strategy used a posteriori. The pairwise difference distribution agreed well with the three-parameter model proposed by Rogers & Harpending (1992). The parameter estimates found for the Basques implied that a demographic expansion (or perhaps two, given the bimodal shape of the distribution) took place sometime between 14500 and 42000 BP which is in agreement with archaeological data. Our sample was compared to other populations for which D-loop sequences were available through the Nei & Miller (1990) distance. In a neighbour-joining tree, all the Caucasoid samples, including the Basques, appeared tightly clustered, whereas African samples were the most distant to the Caucasoids and also the most heterogeneous. Although classical markers, such as blood groups and protein polymorphisms, clearly separate the Basques (and the Sardinians) from other European populations, this distinctiveness was not found using D-loop sequences.