Ireland's position on the western edge of Europe suggests that the genetics of its population should have been relatively undisturbed by the demographic movements that have shaped variation on the mainland. We have typed 221 Y chromosomes from Irish males for seven (slowly evolving) biallelic and six (quickly evolving) simple tandem-repeat markers. When these samples are partitioned by surname, we find significant differences in genetic frequency between those of Irish Gaelic and of foreign origin, and also between those of eastern and western Irish origin. Connaught, the westernmost Irish province, lies at the geographical and genetic extreme of a Europe-wide cline.
Surnames have been used in Ireland from about ad 950 as markers of complex local kinship systems. As both surnames and Y chromosomes are paternally inherited, we divided our Irish sample into seven surname cohorts for which ancient geographical information is known, with some error. Four are of prehistoric, Gaelic origin (Ulster, Munster, Leinster and Connaught) and three are diagnostic of historical influx (Scottish, Norman/Norse and English)1.
The biallelic markers (SRY-1532, M9, YAP, SRY-2627 (ref. 2); SRY-8299 (ref. 3); sY81 (ref. 4); and 92R7 (ref. 5)) define nine haplogroups (clusters of genetic variants) which are highly non-randomly distributed among human populations6, including our samples. In particular, haplogroup 1 (hg 1) has a very high frequency in Ireland (78.1% in the island as a whole).
Surname subdivision reveals a cline in Irish samples, with exogenous samples clearly showing lower frequencies (English, 62.5%; Scottish, 52.9%; Norman/Norse, 83.0%) than Gaelic Irish samples (Leinster, 73.3%; Ulster, 81.1%; Munster, 94.6%), which almost reach fixation in the westernmost province (Connaught, 98.3%). These highly significant differences in the frequency of hg 1 between Irish Gaelic and non-Gaelic Y chromosomes (P<0.001 ) and between eastern and western Gaelic Y chromosomes (P<0.001 ) persist when duplicated surnames are removed.
Eighty per cent (n = 26; ref. 7) of European hg 1 Y chromosomes belong to 'haplotype 15', defined by using the complex p49f/TaqI polymorphic system8. Using this relationship, we estimated that hg 1 frequencies follow a cline within Europe9, extending from the Near East (1.8% in Turkey) to a peak in the Spanish Basque country (89%; ref. 10) in the west (Fig. 1). This cline mirrors other genetic gradients in Europe and is best explained by the migration of Neolithic farmers from the Near East9. When the surname-divided Irish data are appended to this cline, it continues to the western edge of Europe, with hg 1 — the putative pre-Neolithic western European variant — reaching its highest frequency in Connaught (98.3%).
Figure 1: Distribution of observed and estimated haplogroup 1 Y-chromosome haplotypes in Europe.
A cline stretches from a frequency of 1.8% in Turkey to peaks in the Basque country (89%) and the west of Ireland (98% in Connaught, the westernmost marker).
High resolution image and legend (51K)In a maximum-parsimony phylogenetic analysis of both bialellic and simple tandem-repeat (STR) variation between Irish Gaelic haplotypes ( Fig. 2), the hg 1 chromosomes cluster together tightly, with the highest-frequency haplotypes occupying central positions, suggesting a coherent common ancestry. The smaller number of non-hg 1 haplotypes shows no such coherence, consistent with their being immigrants. Their concentration in the eastern Gaelic cohorts may be indicative of a prehistoric influx or of later gene flow across the linguistic barrier from historical migrant groups.
Figure 2: Consensus maximum parsimony networks of Irish Gaelic haplotypes summarizing both single tandem-repeat (DYS19, DYS3891, DYS390, DYS391, DYS392 and DYS393 ; ref. 14) and biallelic variation.
Branch lengths are proportional to the number of mutational steps and node areas are proportional to haplotype frequencies. Haplogroup (hg) 1, blue; hg 2, yellow; hg 21, green; hg 26, white. Asterisk, estimated ancestral haplotype. The separate clustering of haplogroups and the tight clustering of the Irish hg 1 haplotypes around a few numerous, central variants are constant through all most-parsimonious trees and were resistant to repetition of the analysis with randomized inputs.
High resolution image and legend (8K)These findings suggest that hg 1 is the earlier, indigenous Irish variant. By taking the ancestral haplotype as that with the most common allele for each STR and calculating the average squared distance11 (assuming a generation time of 27 years and a mutation rate of 0.21%; ref. 12) between it and all variants (Fig. 2 ), we estimate a date for Irish hg 1 coalescence of 4,200 bp (95% c.i. 1,800–14,800 bp). This relatively recent date (a global estimate of hg 1 coalescence is 30,000 bp; ref. 13) falls well within Ireland's 9,000-year history of human habitation. Although error margins are considerable and include uncertainty related to method, this also provides an upper bound for any agriculturally facilitated population expansion, which, at the fringe of Europe, may have taken place in an insular Mesolithic population of hg 1 genotype.