Classical sickle beta-globin haplotypes exhibit a high degree of long-range haplotype similarity in African and Afro-Caribbean populations

Abstract

Background: The sickle (betas) mutation in the beta-globin gene (HBB) occurs on five "classical" betas haplotype backgrounds in ethnic groups of African ancestry. Strong selection in favour of the betas allele - a consequence of protection from severe malarial infection afforded by heterozygotes - has been associated with a high degree of extended haplotype similarity. The relationship between classical betas haplotypes and long-range haplotype similarity may have both anthropological and clinical implications, but to date has not been explored. Here we evaluate the haplotype similarity of classical betas haplotypes over 400 kb in population samples from Jamaica, The Gambia, and among the Yoruba of Nigeria (Hapmap YRI).

Results: The most common betas sub-haplotype among Jamaicans and the Yoruba was the Benin haplotype, while in The Gambia the Senegal haplotype was observed most commonly. Both subtypes exhibited a high degree of long-range haplotype similarity extending across approximately 400 kb in all three populations. This long-range similarity was significantly greater than that seen for other haplotypes sampled in these populations (P < 0.001), and was independent of marker choice and marker density. Among the Yoruba, Benin haplotypes were highly conserved, with very strong linkage disequilibrium (LD) extending a megabase across the betas mutation.

Conclusion: Two different classical betas haplotypes, sampled from different populations, exhibit comparable and extensive long-range haplotype similarity and strong LD. This LD extends across the adjacent recombination hotspot, and is discernable at distances in excess of 400 kb. Although the multi-centric geographic distribution of betas haplotypes indicates strong subdivision among early Holocene sub-Saharan populations, we find no evidence that selective pressures imposed by falciparum malaria varied in intensity or timing between these subpopulations. Our observations also suggest that cis-acting loci, which may influence outcomes in sickle cell disease, could lie considerable distances away from beta-globin.

Figure 1

Classical β^S haplotypes. The figure illustrates restriction fragment length polymorphisms in a 70 kb region around HBB. Five other globin synthesis genes are shown along with the approximate positions of four RFLP sites used to designate the five classical β^S haplotypes.

Figure 2

Jamaican β^S haplotypes. Haplotypes designated as 'Benin' type are shown in part A; non-Benin type haplotypes are shown in part B. RFLP markers used to distinguish individual β^S haplotypes have shaded marker labels and are outlined by the white border. Haplotypes are arrayed along the Y-axis with SNPs on the X-axis. At each SNP position, the major allele of each SNP is represented in blue and the minor allele in orange. For reference, the 70 kb region of the HBB cluster is indicated at the top of the figure.

Figure 3

Pattern of HBB LD in Yoruba. The figure shows pairwise D' between the β^S/β^A allele (shaded allele) and 180 high frequency SNPs (minor allele frequency > 0.05). D' > 0.9 is shown in red, D' > 0.7 in green, D' > 0.5 in gray, and values < 0.5 in white. The figure also indicates the 70 kb region around β-globin, and the recombination hotspot (dark gray box).

Figure 4

Most common Gambian β^S haplotypes. The 'most common' short-range haplotype, including extension of the haplotype to 400 kb is shown in part A. Individual haplotypes are arrayed along the Y-axis with SNPs on the X-axis. At each SNP position, the major allele of each SNP is represented in blue and the minor allele in orange. The 70 kb region defining the 'short-range' β^S haplotypes is indicated above the figure and by the white border. A comparison of this 70 kb region in Jamaica, Gambia and Yoruba is shown in part B using markers successfully genotyped in all three populations.